[BUG] Dictation stops detecting microphone input after 3 seconds

[Defake]

Describe the bug

I can't use dictation since version 1.6.7 – it consistently stops detecting microphone input after 3-4 seconds.
1.6.8 still has his bug. Only 1.6.6 works for me.
Model: doesn't seem to matter. Parakeet TDT v3 (Multilingual) and Cohere Transcribe both react the same way.

Affected app versions: 1.6.7 and 1.6.8.

Reproduction steps

1. Press shortcut for dictation
2. Start dictating – the waveform reacts and words appear in the preview box as intended
3. After 3-4 seconds the waveform stops reacting, the preview box doesn't show any new words
4. Finish dictation – only the words said during the first 3-4 seconds were captured.

Also model loading didn't affect to the cut off time:

1. I set another model for the Voice Engine
2. Pressed shortcut for dictation
3. The preview box showed that the model was loading, but waveform reacted
4. After 2-3 seconds the model was loaded
5. Within 1 second the waveform stopped reacting – exactly the 3-4 seconds since dictation start. Model loading didn't affect this time.

Expected behavior

Expected to behave like 1.6.6 and earlier – the words to be captured for the whole duration of the dictation.

Actual behavior

It stops detecting the microphone input after 3-4 seconds.

App Version

1.6.8

macOS Version

Tahoe 26.5.2
Apple M1 Pro
Memory | 32 GB

Architecture

Apple Silicon

Logs or crash report

Screenshots or screen recording

No response

[altic-dev]

GIve me more details on your microphone settings, what mic do you use and other details please.

meanwhile - try this n tell me if it works!

#844 (comment)

[Defake]

@altic-dev thanks, reproduced the same issue on 1.6.9 v1. Here're the logs:

Details

[RUN] PID=54768 AppVersion=1.6.9 Build=20 OS=Version 26.5.2 (Build 25F84)
[RUN] Processed by com.FluidApp.app
[17:47:20.336] [INFO] [AppServices] 📦 AppServices singleton created (services not yet initialized)
[17:47:20.661] [INFO] [AppServices] 🎤 Lazily creating ASRService
[17:47:20.662] [INFO] [AppServices] 🔊 Lazily creating AudioHardwareObserver
[17:47:20.843] [DEBUG] [GlobalHotkeyManager] Starting delayed initialization...
[17:47:20.843] [INFO] [GlobalHotkeyManager] Hotkey activation mode set to Toggle
[17:47:20.887] [DEBUG] [ContentView] Accessibility enabled, reinitializing hotkey manager
[17:47:20.887] [INFO] [GlobalHotkeyManager] Manual reinitialization requested
[17:47:20.887] [DEBUG] [GlobalHotkeyManager] Starting delayed initialization...
[17:47:20.888] [DEBUG] [ContentView] Mode transition: nil → Optional(FluidVoice.SidebarItem.welcome)
[17:47:20.888] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2541451.150729 manager set_mode mode=Dictation
[17:47:20.888] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2541451.150786 notch set_mode mode=Dictation
[17:47:21.081] [INFO] [AppDelegate] Application launched [loginItemLaunch=false]
[17:47:21.174] [INFO] [SettingsStore] Initializing app with dock visibility: true
[17:47:21.174] [INFO] [LocalAPIServer] Local API disabled
[17:47:21.176] [INFO] [AppDelegate] Scheduling automatic update check...
[17:47:21.399] [DEBUG] [AppDelegate] Brought main window to front
[17:47:21.405] [DEBUG] [GlobalHotkeyManager] Setup attempt 1/5
[17:47:21.449] [INFO] [GlobalHotkeyManager] Event tap successfully created and enabled
[17:47:21.449] [INFO] [GlobalHotkeyManager] Successfully initialized on attempt 1
[17:47:22.428] [INFO] [ContentView] 🚦 Startup delay complete, signaling UI ready...
[17:47:22.428] [INFO] [AppServices] 🚦 UI Ready signal received - services can now initialize
[17:47:22.428] [DEBUG] [ContentView] Skipping auto-preload - models downloaded manually via UI
[17:47:22.916] [DEBUG] [GlobalHotkeyManager] Setup attempt 1/5
[17:47:22.916] [INFO] [GlobalHotkeyManager] Event tap successfully created and enabled
[17:47:22.916] [INFO] [GlobalHotkeyManager] Successfully initialized on attempt 1
[17:47:23.453] [INFO] [ContentView] 🔊 Starting gated audio initialization...
[17:47:23.579] [INFO] [AudioHardwareObserver] Clamshell state listener registered (closed=false)
[17:47:23.585] [DEBUG] [ASRService] Models exist on disk: true
[17:47:23.735] [DEBUG] [ASRService] Device list change listener registered
[INFO] [DirectCoreAudioLifecycle] t=2541454.009845 Direct capture device resolved off-main reason=prepare:startup device='MacBook Pro Microphone' id=110 elapsedMs=4
[INFO] [DirectCoreAudioLifecycle] t=2541454.018920 Direct capture prepare begin generation=1 device='MacBook Pro Microphone' reason=startup fingerprint={device=110 stream=111 virtual=48000Hz/1ch/32bit/flags=0x9/bytesPerFrame=4 physical=unavailable buffers=[1] nominalHz=48000 bufferFrames=512 variableMax=fixed dataSource=unavailable}
[DEBUG] [DirectCoreAudioLifecycle] t=2541454.034979 Direct capture listener unavailable generation=1 property=device_changed
[DEBUG] [DirectCoreAudioLifecycle] t=2541454.035196 Direct capture listener unavailable generation=1 property=variable_buffer_frame_size
[INFO] [DirectCoreAudioLifecycle] t=2541454.035746 Direct capture prepare end generation=1 elapsedMs=17 fingerprint={device=110 stream=111 virtual=48000Hz/1ch/32bit/flags=0x9/bytesPerFrame=4 physical=unavailable buffers=[1] nominalHz=48000 bufferFrames=512 variableMax=fixed dataSource=unavailable}
[17:47:23.773] [INFO] [ASRService] Prepared direct Core Audio input 'MacBook Pro Microphone' (48000Hz, 512 frames, generation=1, reason=startup)
[17:47:23.773] [INFO] [ASRBenchmark] ASR_BENCH t=2541454.035788 session=0 direct_audio_prewarm reason=startup elapsedMs=38
[17:47:23.773] [INFO] [ContentView] ✅ Audio subsystems initialized
[17:47:23.774] [DEBUG] [ASRService] Models exist on disk: true
[17:47:23.774] [INFO] [ASRService] Models found on disk, auto-loading...
[17:47:23.775] [INFO] [ASRService] ASRService: Created FluidAudio provider [vocabBoosting=false]
[17:47:23.775] [INFO] [ASRService] ensureAsrReady() requested for model=parakeet-tdt [supportsStreaming=true] provider=parakeet-tdt:FluidAudioProvider:FluidAudio (Apple Silicon Optimized)
[17:47:23.856] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2541454.118668 bottom_prepared
[17:47:23.856] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2541454.119104 manager handle_state isRunning=false overlayVisible=false processing=false mode=Dictation
[17:47:23.856] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2541454.119122 manager handle_state_return reason=visibility_unchanged
[17:47:23.856] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2541454.119133 notch text_update status=empty
[17:47:23.929] [DEBUG] [ASRService] ensureAsrReady(begin): provider=FluidAudio (Apple Silicon Optimized), providerReady=false, isAsrReady=false, isRunning=false
[17:47:23.931] [INFO] [ASRService] === ASR INITIALIZATION START ===
[17:47:23.931] [INFO] [ASRService] Using provider: FluidAudio (Apple Silicon Optimized) [providerReady=false]
[17:47:23.931] [INFO] [ASRService] Models already cached on disk: true
[17:47:23.931] [DEBUG] [ASRService] Model cache lookup complete in 0.000s
[17:47:23.931] [INFO] [ASRService] 📦 LOADING cached model into memory...
[17:47:23.931] [INFO] [ASRService] Calling transcriptionProvider.prepare()...
[17:47:23.931] [INFO] [FluidAudioProvider] FluidAudioProvider: Starting model preparation for Parakeet TDT v3 (Multilingual) [version=v3]
[17:47:23.931] [DEBUG] [FluidAudioProvider] FluidAudioProvider: target cache directory=/Users/myuser/Library/Application Support/FluidAudio/Models
[17:47:24.295] [DEBUG] [FluidAudioProvider] FluidAudioProvider: Models downloadAndLoad returned in 0.36s
[17:47:24.295] [DEBUG] [FluidAudioProvider] FluidAudioProvider: Streaming AsrManager initialized
[17:47:24.295] [DEBUG] [FluidAudioProvider] FluidAudioProvider: Word boosting disabled by configuration
[17:47:24.295] [INFO] [FluidAudioProvider] FluidAudioProvider: Models ready [isWordBoostingActive=false, terms=0]
[17:47:24.295] [INFO] [ASRService] ASRService: Provider 'FluidAudio (Apple Silicon Optimized)' prepared successfully in 0.36s
[17:47:24.295] [INFO] [ASRService] ✓ Provider preparation completed in 0.4 seconds
[17:47:24.295] [INFO] [ASRService] ⏳ Model loaded, waiting for first transcription to complete...
[17:47:24.295] [INFO] [ASRService] === ASR INITIALIZATION COMPLETE ===
[17:47:24.295] [INFO] [ASRService] Total initialization time: 0.4 seconds
[17:47:24.358] [INFO] [ASRService] Models auto-loaded successfully on startup
[17:47:24.358] [DEBUG] [ASRService] Audio capture prewarm skipped - backend already prepared
[17:47:24.443] [DEBUG] [ContentView] Initial hotkey manager health check: true
[17:47:24.443] [INFO] [AppDelegate] Performing automatic update check for altic-dev/Fluid-oss
[17:47:33.185] [DEBUG] [ContentView] Mode transition: Optional(FluidVoice.SidebarItem.welcome) → Optional(FluidVoice.SidebarItem.voiceEngine)
[17:47:33.185] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2541463.448273 manager set_mode mode=Dictation
[17:47:33.185] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2541463.448317 notch set_mode mode=Dictation
[17:47:33.301] [DEBUG] [ASRService] Models exist on disk: true
[17:47:39.845] [INFO] [GlobalHotkeyManager] Hotkey route | pressed=dictate(mod) | active=none | asrRunning=false | action=start
[17:47:39.858] [INFO] [GlobalHotkeyManager] Dictate mode hotkey triggered
[17:47:39.859] [DEBUG] [GlobalHotkeyManager] GlobalHotkeyManager: dictate callback path, isRunning=false, isReady=true, model=Parakeet TDT v3 (Multilingual)
[17:47:39.859] [DEBUG] [GlobalHotkeyManager] GlobalHotkeyManager: invoking dictationModeCallback
[17:47:39.859] [INFO] [ContentView] Dictate mode triggered
[17:47:39.859] [DEBUG] [ContentView] ContentView: selected model for dictate hotkey=Parakeet TDT v3 (Multilingual)
[17:47:39.859] [DEBUG] [ContentView] Begin dictation recording for slot primary
[17:47:39.859] [INFO] [AppBenchmark] APP_BENCH t=2541470.108468 begin_recording slot=primary mode=dictate
[17:47:39.859] [INFO] [AppBenchmark] APP_BENCH t=2541470.114836 overlay_mode_request mode=Dictation
[17:47:39.859] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2541470.114867 manager set_mode mode=Dictation
[17:47:39.859] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2541470.114908 notch set_mode mode=Dictation
[17:47:39.859] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2541470.115041 manager instant_show_request mode=Dictation
[17:47:39.859] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2541470.115053 manager show_workitem_execute mode=Dictation
[17:47:39.859] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2541470.115304 notch show_called mode=Dictation state=idle commandExpanded=false
[17:47:39.859] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2541470.115318 notch show_internal_enter mode=Dictation state=idle
[17:47:39.912] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2541470.174833 notch show_internal_route target=bottom
[17:47:39.912] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2541470.174892 notch bottom_route_start mode=Dictation
[17:47:39.912] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2541470.174905 bottom_show_start mode=Dictation windowExists=true
[17:47:39.978] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2541470.241584 bottom_order_front elapsedMs=67
[17:47:39.978] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2541470.241616 bottom_visible elapsedMs=67
[17:47:39.978] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2541470.241703 notch bottom_route_return elapsedMs=67
[17:47:39.979] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2541470.241716 manager show_workitem_return mode=Dictation
[17:47:39.979] [INFO] [AppBenchmark] APP_BENCH t=2541470.241721 overlay_mode_requested mode=Dictation
[17:47:39.979] [INFO] [AppBenchmark] APP_BENCH t=2541470.241725 overlay_phase phase=connecting
[17:47:39.980] [INFO] [AppBenchmark] APP_BENCH t=2541470.242758 asr_start_call
[17:47:39.980] [INFO] [ASRService] 🎤 START() called - beginning recording session
[17:47:39.980] [INFO] [ASRBenchmark] ASR_BENCH t=2541470.242900 session=0 route_recovery_handoff_end generations=[] elapsedMs=0 capturePhase=prepared
[17:47:39.980] [DEBUG] [ASRService] 🧹 Clearing buffers and state
[17:47:39.980] [INFO] [ASRBenchmark] ASR_BENCH t=2541470.242981 session=1 recording_start model=parakeet-tdt provider=nvidia supportsStreaming=true
[17:47:39.980] [DEBUG] [ASRService] ✅ Buffers cleared
[INFO] [DirectCoreAudioLifecycle] t=2541470.261190 Direct capture device resolved off-main reason=recording_start device='MacBook Pro Microphone' id=110 elapsedMs=5
[17:47:40.044] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2541470.307242 notch text_update status=empty
[DEBUG] [DirectCoreAudioLifecycle] t=2541470.307932 Direct capture prepare reuse generation=1 reason=recording_start fingerprint={device=110 stream=111 virtual=48000Hz/1ch/32bit/flags=0x9/bytesPerFrame=4 physical=unavailable buffers=[1] nominalHz=48000 bufferFrames=512 variableMax=fixed dataSource=unavailable}
[INFO] [DirectCoreAudioLifecycle] t=2541470.308426 Direct capture start begin generation=1 device='MacBook Pro Microphone' validationMs=0
[INFO] [DirectCoreAudioLifecycle] t=2541470.366169 Direct capture start end generation=1 elapsedMs=58 fingerprint={device=110 stream=111 virtual=48000Hz/1ch/32bit/flags=0x9/bytesPerFrame=4 physical=unavailable buffers=[1] nominalHz=48000 bufferFrames=512 variableMax=fixed dataSource=unavailable}
[17:47:40.103] [INFO] [ASRBenchmark] ASR_BENCH t=2541470.366280 session=1 audio_backend kind=direct_core_audio device=110 generation=1 frames=512 sampleRate=48000 callbackMs=11
[17:47:40.103] [INFO] [ASRBenchmark] ASR_BENCH t=2541470.366301 session=1 first_pcm_wait_begin attempt=1 attemptID=1 timeoutMs=2000 routeGeneration=0 captureGeneration=1
[17:47:40.111] [INFO] [ASRBenchmark] ASR_BENCH t=2541470.373669 session=1 attempt=1 first_audio sampleCount=171 frameLength=512 sampleRate=48000 bufferMs=11 acquisitionMs=118 elapsedMs=131
[17:47:40.111] [DEBUG] [MicrophonePreferenceCoordinator] Confirmed active microphone as 'MacBook Pro Microphone'
[17:47:40.111] [INFO] [ASRBenchmark] ASR_BENCH t=2541470.373723 session=1 first_pcm_wait_end result=ready attempt=1 attemptID=1 bufferedSamples=171
[17:47:40.111] [INFO] [ASRService] ✅ Audio capture running after first PCM (session=1)
[17:47:40.111] [INFO] [AppBenchmark] APP_BENCH t=2541470.373815 sound_play_request sound=FV_start_0
[17:47:40.194] [INFO] [AppBenchmark] APP_BENCH t=2541470.457377 sound_play_dispatched sound=FV_start_0 elapsedMs=84
[17:47:40.212] [DEBUG] [ContentView] Captured recording app context: app=Arc, bundleId=company.thebrowser.Browser, title=
[17:47:40.235] [DEBUG] [ContentView] Captured preceding text for continuous dictation (chars=0)
[17:47:40.242] [INFO] [AppBenchmark] APP_BENCH t=2541470.505535 overlay_phase phase=recording trigger=first_pcm
[17:47:40.308] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2541470.571535 manager handle_state isRunning=true overlayVisible=true processing=false mode=Dictation
[17:47:40.308] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2541470.571555 manager handle_state_return reason=visibility_unchanged
[17:47:40.308] [DEBUG] [ContentView] ContentView: AI dictation prewarm started slot=primary
[17:47:40.357] [DEBUG] [MediaPlaybackService] MediaPlaybackService: Track info received
- App: Arc
- Bundle: company.thebrowser.Browser
- Title: A Movie (2020)
- isPlaying: false
- playbackRate: 0.0
- Determined playing: false
[17:47:40.357] [DEBUG] [MediaPlaybackService] MediaPlaybackService: Media is not playing, no action needed
[17:47:40.383] [DEBUG] [ASRService] 📡 Starting streaming transcription...
[17:47:40.383] [INFO] [ASRBenchmark] ASR_BENCH t=2541470.646578 session=1 streaming_timer_start intervalMs=600 minSamples=16000
[17:47:40.383] [DEBUG] [ASRService] Starting streaming transcription task (interval: 0.6s, minSamples: 16000)
[17:47:40.383] [INFO] [ASRService] ✅ START() completed successfully
[17:47:40.384] [INFO] [AppBenchmark] APP_BENCH t=2541470.646699 asr_start_return elapsedMs=404
[17:47:40.431] [DEBUG] [ASRService] 🔄 runStreamingLoop() - ENTERED
[17:47:40.431] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[17:47:40.431] [DEBUG] [ASRService] Waiting for more audio data (5291/16000 samples)
[17:47:40.431] [INFO] [ASRBenchmark] ASR_BENCH t=2541470.694479 session=1 chunk_wait index=1 ageMs=452 samples=5291 minSamples=16000
[17:47:40.431] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[17:47:40.477] [INFO] [PrivateAIProvider] Private provider creating runtime model=fluid-1 backend=mlx:/Volumes/FluidVoice Installer/FluidVoice.app/Contents/Helpers/fluid-intelligence-mlx:/Users/myuser/Library/Application Support/FluidIntelligence/Models/fluid-1-nvfp4-mlx:mtp:/Users/myuser/Library/Application Support/FluidIntelligence/Models/gemma-4-E2B-it-qat-assistant-bf16-mlx-mtp previous=none
[17:47:41.055] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[17:47:41.055] [DEBUG] [ASRService] Waiting for more audio data (15190/16000 samples)
[17:47:41.055] [INFO] [ASRBenchmark] ASR_BENCH t=2541471.317882 session=1 chunk_wait index=2 ageMs=1075 samples=15190 minSamples=16000
[17:47:41.055] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[17:47:41.697] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[17:47:41.697] [INFO] [ASRBenchmark] ASR_BENCH t=2541471.960594 session=1 chunk_start index=3 ageMs=1718 samples=25430 newSamples=25430 audioMs=1589 provider=FluidAudio (Apple Silicon Optimized)
[17:47:41.698] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 25430) using FluidAudio (Apple Silicon Optimized)
[17:47:42.327] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=25430 audioMs=1589 elapsedMs=629 textChars=0 rtf=0.396
[17:47:42.329] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.63s
[17:47:42.330] [INFO] [ASRBenchmark] ASR_BENCH t=2541472.592820 session=1 chunk_done index=3 elapsedMs=632 ageMs=2350 samples=25430 rawChars=0 cleanedChars=0 rtf=0.396
[17:47:42.330] [DEBUG] [ASRService] ⚠️ Transcription slow (0.63s > 0.6s), skipping next chunk
[17:47:42.330] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[17:47:42.341] [INFO] [ASRService] ✅ Model warmed up - first streaming transcription completed
[17:47:42.975] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[17:47:42.975] [DEBUG] [ASRService] ⚠️ Skipping chunk for ANE recovery
[17:47:42.975] [INFO] [ASRBenchmark] ASR_BENCH t=2541473.238252 session=1 chunk_skip index=4 reason=recovery ageMs=2995
[17:47:42.975] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[17:47:43.582] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[17:47:43.582] [INFO] [ASRBenchmark] ASR_BENCH t=2541473.845520 session=1 chunk_start index=5 ageMs=3603 samples=55638 newSamples=30208 audioMs=3477 provider=FluidAudio (Apple Silicon Optimized)
[17:47:43.582] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 55638) using FluidAudio (Apple Silicon Optimized)
[17:47:43.668] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=55638 audioMs=3477 elapsedMs=85 textChars=46 rtf=0.024
[17:47:43.668] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.09s
[17:47:43.668] [DEBUG] [ASRService] ✅ Streaming: 'One, two, three, four, five, six, seven, eight' (0.09s)
[17:47:43.668] [INFO] [ASRBenchmark] ASR_BENCH t=2541473.931250 session=1 chunk_done index=5 elapsedMs=86 ageMs=3688 samples=55638 rawChars=46 cleanedChars=46 rtf=0.025
[17:47:43.668] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[17:47:44.268] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[17:47:44.268] [INFO] [ASRBenchmark] ASR_BENCH t=2541474.531436 session=1 chunk_start index=6 ageMs=4288 samples=66560 newSamples=10922 audioMs=4160 provider=FluidAudio (Apple Silicon Optimized)
[17:47:44.268] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 66560) using FluidAudio (Apple Silicon Optimized)
[17:47:44.362] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=66560 audioMs=4160 elapsedMs=94 textChars=46 rtf=0.023
[17:47:44.362] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.09s
[17:47:44.362] [DEBUG] [ASRService] ✅ Streaming: 'One, two, three, four, five, six, seven, eight' (0.09s)
[17:47:44.362] [INFO] [ASRBenchmark] ASR_BENCH t=2541474.625455 session=1 chunk_done index=6 elapsedMs=94 ageMs=4382 samples=66560 rawChars=46 cleanedChars=46 rtf=0.023
[17:47:44.362] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[17:47:45.001] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[17:47:45.001] [INFO] [ASRBenchmark] ASR_BENCH t=2541475.263751 session=1 chunk_start index=7 ageMs=5021 samples=78336 newSamples=11776 audioMs=4896 provider=FluidAudio (Apple Silicon Optimized)
[17:47:45.001] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 78336) using FluidAudio (Apple Silicon Optimized)
[17:47:45.077] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=78336 audioMs=4896 elapsedMs=76 textChars=46 rtf=0.016
[17:47:45.077] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.08s
[17:47:45.077] [DEBUG] [ASRService] ✅ Streaming: 'One, two, three, four, five, six, seven, eight' (0.08s)
[17:47:45.077] [INFO] [ASRBenchmark] ASR_BENCH t=2541475.340592 session=1 chunk_done index=7 elapsedMs=77 ageMs=5098 samples=78336 rawChars=46 cleanedChars=46 rtf=0.016
[17:47:45.077] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[17:47:45.720] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[17:47:45.720] [INFO] [ASRBenchmark] ASR_BENCH t=2541475.983512 session=1 chunk_start index=8 ageMs=5741 samples=89771 newSamples=11435 audioMs=5611 provider=FluidAudio (Apple Silicon Optimized)
[17:47:45.720] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 89771) using FluidAudio (Apple Silicon Optimized)
[17:47:45.801] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=89771 audioMs=5611 elapsedMs=81 textChars=46 rtf=0.014
[17:47:45.801] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.08s
[17:47:45.802] [DEBUG] [ASRService] ✅ Streaming: 'One, two, three, four, five, six, seven, eight' (0.08s)
[17:47:45.802] [INFO] [ASRBenchmark] ASR_BENCH t=2541476.064877 session=1 chunk_done index=8 elapsedMs=81 ageMs=5822 samples=89771 rawChars=46 cleanedChars=46 rtf=0.014
[17:47:45.802] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[17:47:46.442] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[17:47:46.442] [INFO] [ASRBenchmark] ASR_BENCH t=2541476.705135 session=1 chunk_start index=9 ageMs=6462 samples=101376 newSamples=11605 audioMs=6336 provider=FluidAudio (Apple Silicon Optimized)
[17:47:46.442] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 101376) using FluidAudio (Apple Silicon Optimized)
[17:47:46.496] [DEBUG] [ContentView] AI dictation prewarm complete slot=primary
[17:47:46.516] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=101376 audioMs=6336 elapsedMs=74 textChars=46 rtf=0.012
[17:47:46.516] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.07s
[17:47:46.516] [DEBUG] [ASRService] ✅ Streaming: 'One, two, three, four, five, six, seven, eight' (0.07s)
[17:47:46.516] [INFO] [ASRBenchmark] ASR_BENCH t=2541476.779556 session=1 chunk_done index=9 elapsedMs=74 ageMs=6537 samples=101376 rawChars=46 cleanedChars=46 rtf=0.012
[17:47:46.516] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[17:47:47.156] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[17:47:47.156] [INFO] [ASRBenchmark] ASR_BENCH t=2541477.419063 session=1 chunk_start index=10 ageMs=7176 samples=112811 newSamples=11435 audioMs=7051 provider=FluidAudio (Apple Silicon Optimized)
[17:47:47.156] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 112811) using FluidAudio (Apple Silicon Optimized)
[17:47:47.243] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=112811 audioMs=7051 elapsedMs=87 textChars=47 rtf=0.012
[17:47:47.243] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.09s
[17:47:47.243] [DEBUG] [ASRService] ✅ Streaming: 'One, two, three, four, five, six, seven, eight.' (0.09s)
[17:47:47.243] [INFO] [ASRBenchmark] ASR_BENCH t=2541477.506550 session=1 chunk_done index=10 elapsedMs=87 ageMs=7264 samples=112811 rawChars=47 cleanedChars=47 rtf=0.012
[17:47:47.243] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[17:47:47.850] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[17:47:47.850] [INFO] [ASRBenchmark] ASR_BENCH t=2541478.112922 session=1 chunk_start index=11 ageMs=7870 samples=123904 newSamples=11093 audioMs=7744 provider=FluidAudio (Apple Silicon Optimized)
[17:47:47.850] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 123904) using FluidAudio (Apple Silicon Optimized)
[17:47:47.925] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=123904 audioMs=7744 elapsedMs=75 textChars=47 rtf=0.010
[17:47:47.925] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.08s
[17:47:47.925] [DEBUG] [ASRService] ✅ Streaming: 'One, two, three, four, five, six, seven, eight.' (0.08s)
[17:47:47.925] [INFO] [ASRBenchmark] ASR_BENCH t=2541478.188248 session=1 chunk_done index=11 elapsedMs=75 ageMs=7945 samples=123904 rawChars=47 cleanedChars=47 rtf=0.010
[17:47:47.925] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[17:47:47.987] [INFO] [GlobalHotkeyManager] Cancel shortcut pressed - cancelling recording
[17:47:47.987] [INFO] [GlobalHotkeyManager] Cancel shortcut pressed - cancel callback handled
[17:47:47.993] [INFO] [ASRService] 🛑 Stopping recording - releasing audio devices
[INFO] [DirectCoreAudioLifecycle] t=2541478.256594 Direct capture stop begin generation=1 reason=stop_without_transcription retainPrepared=false
[INFO] [DirectCoreAudioLifecycle] t=2541478.273743 Direct capture invalidate begin generation=1 reason=stop_without_transcription
[DEBUG] [DirectCoreAudioLifecycle] t=2541478.273828 Direct capture listener drain generation=1 count=11 elapsedMs=0
[INFO] [DirectCoreAudioLifecycle] t=2541478.274152 Direct capture invalidate end generation=1 elapsedMs=0 reason=stop_without_transcription
[INFO] [DirectCoreAudioLifecycle] t=2541478.274170 Direct capture stop end generation=1 elapsedMs=18 status=0 droppedPackets=0 retained=false
[17:47:48.047] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2541478.310220 manager handle_state isRunning=false overlayVisible=true processing=false mode=Dictation
[17:47:48.047] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2541478.310246 manager hide_request delayMs=30
[17:47:48.047] [DEBUG] [ASRService] Audio capture stopped
[17:47:48.047] [DEBUG] [ASRService] Audio engine retired (stop_without_transcription)
[17:47:48.047] [INFO] [ASRBenchmark] ASR_BENCH t=2541478.310426 session=1 streaming_timer_stop begin
[17:47:48.158] [DEBUG] [ASRService] Streaming transcription task cancelled
[17:47:48.158] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2541478.421182 manager hide_workitem_execute
[17:47:48.158] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2541478.421205 manager hide_workitem_return
[17:47:48.196] [INFO] [ASRBenchmark] ASR_BENCH t=2541478.459653 session=1 streaming_timer_stop end elapsedMs=149 completedChunks=8
[17:47:48.197] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2541478.459792 notch hide_called state=idle bottomVisible=true
[17:47:48.197] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2541478.459907 bottom_hide_start windowExists=true
[17:47:48.197] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2541478.459936 bottom_hide_animation_start
[17:47:48.273] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2541478.535762 notch text_update status=empty
[17:47:48.302] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2541478.565420 bottom_hide_complete elapsedMs=106
[17:47:48.302] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2541478.565460 notch hide_return reason=not_visible state=idle notchExists=false
[17:47:48.302] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2541478.565466 notch hide_complete target=none elapsedMs=106
[17:47:50.136] [INFO] [AppDelegate] Application will terminate
[17:47:50.408] [INFO] [PrivateAIProvider] Private provider unloaded cached runtime model=fluid-1 reason=termination
[17:47:50.408] [INFO] [ASRBenchmark] ASR_BENCH t=2541480.671325 session=1 route_recovery_shutdown elapsedMs=0
[17:47:50.408] [DEBUG] [ASRService] Audio engine retired (app_termination)
[17:47:50.408] [INFO] [ASRBenchmark] ASR_BENCH t=2541480.671428 session=1 audio_engine_shutdown elapsedMs=0
[17:47:50.408] [INFO] [ASRBenchmark] ASR_BENCH t=2541480.671481 session=1 direct_capture_shutdown phase=shutDown elapsedMs=0
[17:47:50.456] [INFO] [AppServices] 🎤 Lazily creating ASRService

I use built-in macbook's microphone. What details do you mean specifically?

Audio settings in FluidVoice:

Details

same in 1.6.9:

[altic-dev]

what mic do you use? it looks like the audio packets are kinda empty - are you whispering or using a mic that's far away or anything?

will you be able to try with some other mic you got? perhaps your phone mic or anything else?

[altic-dev]

1.6.9-v3

run this after dfragging it into apps and let mek now if it works and also the logs :) trying to figure this out for you.

[Defake]

@altic-dev thanks for helping. This version reproduces the issue as well.

As I said, I use built-in macbook's microphone. I'll try an external mic and update my comment.
The microphone sensivity is certainly not the issue, because during the first 3-4 seconds the app is detecting my voice perfectly and captures all words. After those 3-4 seconds, the waveform stops reacting and it doesn't capture anything after that.
In the logs you can see "Model answer (A): 1, 2, 3, 4, 5, 6, 7." – But in reality I dictated from 1 to 15, and after the 7 it stopped hearing me.

The logs:

Details

[RUN] PID=34698 AppVersion=1.6.9 Build=20 OS=Version 26.5.2 (Build 25F84)
[RUN] Processed by com.FluidApp.app
[09:03:25.655] [INFO] [AppServices] 📦 AppServices singleton created (services not yet initialized)
[09:03:25.946] [INFO] [AppServices] 🎤 Lazily creating ASRService
[09:03:25.947] [INFO] [AppServices] 🔊 Lazily creating AudioHardwareObserver
[09:03:26.250] [DEBUG] [GlobalHotkeyManager] Starting delayed initialization...
[09:03:26.250] [INFO] [GlobalHotkeyManager] Hotkey activation mode set to Toggle
[09:03:26.336] [DEBUG] [ContentView] Accessibility enabled, reinitializing hotkey manager
[09:03:26.336] [INFO] [GlobalHotkeyManager] Manual reinitialization requested
[09:03:26.336] [DEBUG] [GlobalHotkeyManager] Starting delayed initialization...
[09:03:26.337] [DEBUG] [ContentView] Mode transition: nil → Optional(FluidVoice.SidebarItem.welcome)
[09:03:26.337] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594468.989212 manager set_mode mode=Dictation
[09:03:26.337] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594468.989267 notch set_mode mode=Dictation
[09:03:26.645] [INFO] [AppDelegate] Application launched [loginItemLaunch=false]
[09:03:26.760] [INFO] [SettingsStore] Initializing app with dock visibility: true
[09:03:26.760] [INFO] [LocalAPIServer] Local API disabled
[09:03:26.760] [INFO] [AppDelegate] Scheduling automatic update check...
[09:03:26.991] [DEBUG] [AppDelegate] Brought main window to front
[09:03:27.003] [DEBUG] [GlobalHotkeyManager] Setup attempt 1/5
[09:03:27.073] [INFO] [GlobalHotkeyManager] Event tap successfully created and enabled
[09:03:27.073] [INFO] [GlobalHotkeyManager] Successfully initialized on attempt 1
[09:03:27.730] [INFO] [ContentView] 🚦 Startup delay complete, signaling UI ready...
[09:03:27.736] [INFO] [AppServices] 🚦 UI Ready signal received - services can now initialize
[09:03:28.015] [DEBUG] [ContentView] Skipping auto-preload - models downloaded manually via UI
[09:03:28.582] [DEBUG] [GlobalHotkeyManager] Setup attempt 1/5
[09:03:28.583] [INFO] [GlobalHotkeyManager] Event tap successfully created and enabled
[09:03:28.583] [INFO] [GlobalHotkeyManager] Successfully initialized on attempt 1
[09:03:29.128] [INFO] [ContentView] 🔊 Starting gated audio initialization...
[09:03:29.345] [INFO] [AudioHardwareObserver] Clamshell state listener registered (closed=false)
[09:03:29.346] [DEBUG] [ASRService] Models exist on disk: true
[09:03:29.481] [DEBUG] [ASRService] Device list change listener registered
[INFO] [DirectCoreAudioLifecycle] t=2594472.146350 Direct capture device resolved off-main reason=prepare:startup device='MacBook Pro Microphone' id=110 elapsedMs=4
[INFO] [DirectCoreAudioLifecycle] t=2594472.146972 Direct capture prepare begin generation=1 device='MacBook Pro Microphone' reason=startup fingerprint={device=110 stream=111 virtual=48000Hz/1ch/32bit/flags=0x9/bytesPerFrame=4 physical=unavailable buffers=[1] nominalHz=48000 bufferFrames=512 variableMax=fixed dataSource=unavailable}
[DEBUG] [DirectCoreAudioLifecycle] t=2594472.176460 Direct capture listener unavailable generation=1 property=device_changed
[DEBUG] [DirectCoreAudioLifecycle] t=2594472.176786 Direct capture listener unavailable generation=1 property=variable_buffer_frame_size
[INFO] [DirectCoreAudioLifecycle] t=2594472.177938 Direct capture prepare end generation=1 elapsedMs=31 fingerprint={device=110 stream=111 virtual=48000Hz/1ch/32bit/flags=0x9/bytesPerFrame=4 physical=unavailable buffers=[1] nominalHz=48000 bufferFrames=512 variableMax=fixed dataSource=unavailable}
[09:03:29.542] [INFO] [ASRService] Prepared direct Core Audio input 'MacBook Pro Microphone' (48000Hz, 512 frames, generation=1, reason=startup)
[09:03:29.542] [INFO] [ASRBenchmark] ASR_BENCH t=2594472.195369 session=0 direct_audio_prewarm reason=startup elapsedMs=61
[09:03:29.543] [INFO] [ContentView] ✅ Audio subsystems initialized
[09:03:29.546] [DEBUG] [ASRService] Models exist on disk: true
[09:03:29.547] [INFO] [ASRService] Models found on disk, auto-loading...
[09:03:29.548] [INFO] [ASRService] ASRService: Created FluidAudio provider [vocabBoosting=false]
[09:03:29.548] [INFO] [ASRService] ensureAsrReady() requested for model=parakeet-tdt [supportsStreaming=true] provider=parakeet-tdt:FluidAudioProvider:FluidAudio (Apple Silicon Optimized)
[09:03:29.651] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594472.304291 bottom_prepared
[09:03:29.652] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594472.304651 manager handle_state isRunning=false overlayVisible=false processing=false mode=Dictation
[09:03:29.652] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594472.304675 manager handle_state_return reason=visibility_unchanged
[09:03:29.652] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594472.304688 notch text_update status=empty
[09:03:29.703] [DEBUG] [ASRService] ensureAsrReady(begin): provider=FluidAudio (Apple Silicon Optimized), providerReady=false, isAsrReady=false, isRunning=false
[09:03:29.705] [INFO] [ASRService] === ASR INITIALIZATION START ===
[09:03:29.705] [INFO] [ASRService] Using provider: FluidAudio (Apple Silicon Optimized) [providerReady=false]
[09:03:29.705] [INFO] [ASRService] Models already cached on disk: true
[09:03:29.705] [DEBUG] [ASRService] Model cache lookup complete in 0.000s
[09:03:29.705] [INFO] [ASRService] 📦 LOADING cached model into memory...
[09:03:29.705] [INFO] [ASRService] Calling transcriptionProvider.prepare()...
[09:03:29.705] [INFO] [FluidAudioProvider] FluidAudioProvider: Starting model preparation for Parakeet TDT v3 (Multilingual) [version=v3]
[09:03:29.705] [DEBUG] [FluidAudioProvider] FluidAudioProvider: target cache directory=/Users/myuser/Library/Application Support/FluidAudio/Models
[09:03:30.057] [DEBUG] [ContentView] Initial hotkey manager health check: true
[09:03:30.057] [INFO] [AppDelegate] Performing automatic update check for altic-dev/Fluid-oss
[09:03:30.057] [DEBUG] [FluidAudioProvider] FluidAudioProvider: Models downloadAndLoad returned in 0.35s
[09:03:30.124] [DEBUG] [FluidAudioProvider] FluidAudioProvider: Streaming AsrManager initialized
[09:03:30.124] [DEBUG] [FluidAudioProvider] FluidAudioProvider: Word boosting disabled by configuration
[09:03:30.124] [INFO] [FluidAudioProvider] FluidAudioProvider: Models ready [isWordBoostingActive=false, terms=0]
[09:03:30.124] [INFO] [ASRService] ASRService: Provider 'FluidAudio (Apple Silicon Optimized)' prepared successfully in 0.42s
[09:03:30.124] [INFO] [ASRService] ✓ Provider preparation completed in 0.4 seconds
[09:03:30.124] [INFO] [ASRService] ⏳ Model loaded, waiting for first transcription to complete...
[09:03:30.124] [INFO] [ASRService] === ASR INITIALIZATION COMPLETE ===
[09:03:30.124] [INFO] [ASRService] Total initialization time: 0.4 seconds
[09:03:30.197] [INFO] [ASRService] Models auto-loaded successfully on startup
[09:03:30.198] [DEBUG] [ASRService] Audio capture prewarm skipped - backend already prepared
[09:03:35.690] [INFO] [ContentView] ContentView: startRecording() for model=Parakeet TDT v3 (Multilingual), supportsStreaming=true
[09:03:35.690] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594478.343024 manager set_mode mode=Dictation
[09:03:35.690] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594478.343071 notch set_mode mode=Dictation
[09:03:35.690] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594478.343144 manager instant_show_request mode=Dictation
[09:03:35.690] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594478.343155 manager show_workitem_execute mode=Dictation
[09:03:35.690] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594478.343275 notch show_called mode=Dictation state=idle commandExpanded=false
[09:03:35.690] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594478.343286 notch show_internal_enter mode=Dictation state=idle
[09:03:35.691] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594478.343572 notch show_internal_route target=bottom
[09:03:35.691] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594478.343587 notch bottom_route_start mode=Dictation
[09:03:35.691] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594478.343592 bottom_show_start mode=Dictation windowExists=true
[09:03:35.701] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594478.353970 bottom_order_front elapsedMs=10
[09:03:35.701] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594478.354015 bottom_visible elapsedMs=10
[09:03:35.701] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594478.354071 notch bottom_route_return elapsedMs=10
[09:03:35.701] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594478.354081 manager show_workitem_return mode=Dictation
[09:03:35.701] [INFO] [AppBenchmark] APP_BENCH t=2594478.354085 overlay_phase phase=connecting
[09:03:35.703] [INFO] [ASRService] 🎤 START() called - beginning recording session
[09:03:35.704] [INFO] [ASRBenchmark] ASR_BENCH t=2594478.356597 session=0 route_recovery_handoff_end generations=[] elapsedMs=0 capturePhase=prepared
[09:03:35.704] [DEBUG] [ASRService] 🧹 Clearing buffers and state
[09:03:35.704] [INFO] [ASRBenchmark] ASR_BENCH t=2594478.356690 session=1 recording_start model=parakeet-tdt provider=nvidia supportsStreaming=true
[09:03:35.704] [DEBUG] [ASRService] ✅ Buffers cleared
[09:03:35.716] [DEBUG] [ContentView] ContentView: pre-load model task started
[09:03:35.716] [INFO] [ASRService] ensureAsrReady() requested for model=parakeet-tdt [supportsStreaming=true] provider=parakeet-tdt:FluidAudioProvider:FluidAudio (Apple Silicon Optimized)
[INFO] [DirectCoreAudioLifecycle] t=2594478.373043 Direct capture device resolved off-main reason=recording_start device='MacBook Pro Microphone' id=110 elapsedMs=4
[09:03:35.771] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594478.424265 notch text_update status=empty
[09:03:35.771] [DEBUG] [ASRService] ensureAsrReady(begin): provider=FluidAudio (Apple Silicon Optimized), providerReady=true, isAsrReady=true, isRunning=false
[09:03:35.771] [DEBUG] [ASRService] ASR already ready with loaded models, skipping initialization
[DEBUG] [DirectCoreAudioLifecycle] t=2594478.425130 Direct capture prepare reuse generation=1 reason=recording_start fingerprint={device=110 stream=111 virtual=48000Hz/1ch/32bit/flags=0x9/bytesPerFrame=4 physical=unavailable buffers=[1] nominalHz=48000 bufferFrames=512 variableMax=fixed dataSource=unavailable}
[INFO] [DirectCoreAudioLifecycle] t=2594478.425713 Direct capture start begin generation=1 device='MacBook Pro Microphone' validationMs=1
[INFO] [DirectCoreAudioLifecycle] t=2594478.486306 Direct capture start end generation=1 elapsedMs=61 fingerprint={device=110 stream=111 virtual=48000Hz/1ch/32bit/flags=0x9/bytesPerFrame=4 physical=unavailable buffers=[1] nominalHz=48000 bufferFrames=512 variableMax=fixed dataSource=unavailable}
[09:03:35.851] [DEBUG] [ContentView] Model pre-loaded during recording
[09:03:35.851] [INFO] [ASRBenchmark] ASR_BENCH t=2594478.504257 session=1 audio_backend kind=direct_core_audio device=110 generation=1 frames=512 sampleRate=48000 callbackMs=11
[09:03:35.851] [INFO] [ASRBenchmark] ASR_BENCH t=2594478.504288 session=1 first_pcm_wait_begin attempt=1 attemptID=1 timeoutMs=2000 routeGeneration=0 captureGeneration=1
[09:03:35.851] [INFO] [ASRBenchmark] ASR_BENCH t=2594478.504343 session=1 attempt=1 first_audio sampleCount=171 frameLength=512 sampleRate=48000 bufferMs=11 acquisitionMs=125 elapsedMs=137
[09:03:35.921] [DEBUG] [MicrophonePreferenceCoordinator] Confirmed active microphone as 'MacBook Pro Microphone'
[09:03:35.921] [INFO] [ASRBenchmark] ASR_BENCH t=2594478.573722 session=1 first_pcm_wait_end result=ready attempt=1 attemptID=1 bufferedSamples=1366
[09:03:35.921] [INFO] [ASRService] ✅ Audio capture running after first PCM (session=1)
[09:03:35.921] [INFO] [AppBenchmark] APP_BENCH t=2594478.573832 sound_play_request sound=FV_start_0
[09:03:35.927] [DEBUG] [ContentView] Captured recording app context: app=FluidVoice, bundleId=com.FluidApp.app, title=FluidVoice
[09:03:35.947] [DEBUG] [ContentView] Captured preceding text for continuous dictation (chars=0)
[09:03:35.955] [INFO] [AppBenchmark] APP_BENCH t=2594478.607901 overlay_phase phase=recording trigger=first_pcm
[09:03:36.016] [INFO] [AppBenchmark] APP_BENCH t=2594478.668498 sound_play_dispatched sound=FV_start_0 elapsedMs=95
[09:03:36.073] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594478.725569 manager handle_state isRunning=true overlayVisible=true processing=false mode=Dictation
[09:03:36.073] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594478.725595 manager handle_state_return reason=visibility_unchanged
[09:03:36.073] [DEBUG] [ContentView] ContentView: AI dictation prewarm started slot=primary
[09:03:36.073] [DEBUG] [MediaPlaybackService] MediaPlaybackService: Track info received
- App: Arc
- Bundle: company.thebrowser.Browser
- Title: A Movie
- isPlaying: false
- playbackRate: 0.0
- Determined playing: false
- elapsedMs: 117
[09:03:36.073] [DEBUG] [MediaPlaybackService] MediaPlaybackService: Media is not playing, no action needed (elapsedMs: 117)
[09:03:36.197] [DEBUG] [ASRService] 📡 Starting streaming transcription...
[09:03:36.197] [INFO] [ASRBenchmark] ASR_BENCH t=2594478.849843 session=1 streaming_timer_start intervalMs=600 minSamples=16000
[09:03:36.197] [DEBUG] [ASRService] Starting streaming transcription task (interval: 0.6s, minSamples: 16000)
[09:03:36.197] [INFO] [ASRService] ✅ START() completed successfully
[09:03:36.297] [DEBUG] [ASRService] 🔄 runStreamingLoop() - ENTERED
[09:03:36.297] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[09:03:36.297] [DEBUG] [ASRService] Waiting for more audio data (7339/16000 samples)
[09:03:36.297] [INFO] [ASRBenchmark] ASR_BENCH t=2594478.949656 session=1 chunk_wait index=1 ageMs=593 samples=7339 minSamples=16000
[09:03:36.297] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[09:03:36.612] [INFO] [PrivateAIProvider] Private provider creating runtime model=fluid-1 backend=mlx:/Applications/FluidVoice_169_3.app/Contents/Helpers/fluid-intelligence-mlx:/Users/myuser/Library/Application Support/FluidIntelligence/Models/fluid-1-nvfp4-mlx:mtp:/Users/myuser/Library/Application Support/FluidIntelligence/Models/gemma-4-E2B-it-qat-assistant-bf16-mlx-mtp previous=none
[09:03:36.833] [INFO] [ASRBenchmark] ASR_BENCH t=2594479.486262 session=1 capture_health attempt=1 audioMs=1003 samples=16043 rms=0.014611 peak=0.144299 silent=false inputUID=BuiltInMicrophoneDevice
[09:03:36.938] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[09:03:36.938] [INFO] [ASRBenchmark] ASR_BENCH t=2594479.591211 session=1 chunk_start index=2 ageMs=1235 samples=17579 newSamples=17579 audioMs=1099 provider=FluidAudio (Apple Silicon Optimized)
[09:03:36.939] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 17579) using FluidAudio (Apple Silicon Optimized)
[09:03:37.418] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=17579 audioMs=1099 elapsedMs=480 textChars=0 rtf=0.437
[09:03:37.418] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.48s
[09:03:37.420] [INFO] [ASRBenchmark] ASR_BENCH t=2594480.072651 session=1 chunk_done index=2 elapsedMs=481 ageMs=1716 samples=17579 rawChars=0 cleanedChars=0 rtf=0.437
[09:03:37.420] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[09:03:37.420] [INFO] [ASRService] ✅ Model warmed up - first streaming transcription completed
[09:03:37.841] [INFO] [ASRBenchmark] ASR_BENCH t=2594480.493857 session=1 capture_health attempt=1 audioMs=2005 samples=16043 rms=0.042708 peak=0.233148 silent=false inputUID=BuiltInMicrophoneDevice
[09:03:38.059] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[09:03:38.060] [INFO] [ASRBenchmark] ASR_BENCH t=2594480.712334 session=1 chunk_start index=3 ageMs=2356 samples=35499 newSamples=17920 audioMs=2219 provider=FluidAudio (Apple Silicon Optimized)
[09:03:38.060] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 35499) using FluidAudio (Apple Silicon Optimized)
[09:03:38.304] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=35499 audioMs=2219 elapsedMs=243 textChars=18 rtf=0.110
[09:03:38.305] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.24s
[09:03:38.305] [DEBUG] [ASRService] ✅ Streaming: 'One two three four' (0.24s)
[09:03:38.305] [INFO] [ASRBenchmark] ASR_BENCH t=2594480.956292 session=1 chunk_done index=3 elapsedMs=244 ageMs=2600 samples=35499 rawChars=18 cleanedChars=18 rtf=0.110
[09:03:38.315] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[09:03:38.838] [INFO] [ASRBenchmark] ASR_BENCH t=2594481.491165 session=1 capture_health attempt=1 audioMs=3008 samples=16042 rms=0.029513 peak=0.188823 silent=false inputUID=BuiltInMicrophoneDevice
[09:03:38.944] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[09:03:38.944] [INFO] [ASRBenchmark] ASR_BENCH t=2594481.596818 session=1 chunk_start index=4 ageMs=3240 samples=49664 newSamples=14165 audioMs=3104 provider=FluidAudio (Apple Silicon Optimized)
[09:03:38.944] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 49664) using FluidAudio (Apple Silicon Optimized)
[09:03:39.134] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=49664 audioMs=3104 elapsedMs=191 textChars=28 rtf=0.062
[09:03:39.135] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.19s
[09:03:39.135] [DEBUG] [ASRService] ✅ Streaming: 'One, two, three, four, five.' (0.19s)
[09:03:39.135] [INFO] [ASRBenchmark] ASR_BENCH t=2594481.787939 session=1 chunk_done index=4 elapsedMs=191 ageMs=3431 samples=49664 rawChars=28 cleanedChars=28 rtf=0.061
[09:03:39.135] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[09:03:39.774] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[09:03:39.774] [INFO] [ASRBenchmark] ASR_BENCH t=2594482.426885 session=1 chunk_start index=5 ageMs=4070 samples=62976 newSamples=13312 audioMs=3936 provider=FluidAudio (Apple Silicon Optimized)
[09:03:39.774] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 62976) using FluidAudio (Apple Silicon Optimized)
[09:03:39.841] [INFO] [ASRBenchmark] ASR_BENCH t=2594482.493894 session=1 capture_health attempt=1 audioMs=4011 samples=16043 rms=0.023177 peak=0.173746 silent=false inputUID=BuiltInMicrophoneDevice
[09:03:39.861] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=62976 audioMs=3936 elapsedMs=87 textChars=39 rtf=0.022
[09:03:39.861] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.09s
[09:03:39.862] [DEBUG] [ASRService] ✅ Streaming: 'One, two, three, four, five, six, seven' (0.09s)
[09:03:39.862] [INFO] [ASRBenchmark] ASR_BENCH t=2594482.514641 session=1 chunk_done index=5 elapsedMs=88 ageMs=4158 samples=62976 rawChars=39 cleanedChars=39 rtf=0.022
[09:03:39.862] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[09:03:40.497] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[09:03:40.497] [INFO] [ASRBenchmark] ASR_BENCH t=2594483.150436 session=1 chunk_start index=6 ageMs=4794 samples=74582 newSamples=11606 audioMs=4661 provider=FluidAudio (Apple Silicon Optimized)
[09:03:40.497] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 74582) using FluidAudio (Apple Silicon Optimized)
[09:03:40.584] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=74582 audioMs=4661 elapsedMs=86 textChars=39 rtf=0.018
[09:03:40.584] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.09s
[09:03:40.584] [DEBUG] [ASRService] ✅ Streaming: 'One, two, three, four, five, six, seven' (0.09s)
[09:03:40.584] [INFO] [ASRBenchmark] ASR_BENCH t=2594483.237219 session=1 chunk_done index=6 elapsedMs=87 ageMs=4881 samples=74582 rawChars=39 cleanedChars=39 rtf=0.019
[09:03:40.584] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[09:03:40.844] [INFO] [ASRBenchmark] ASR_BENCH t=2594483.496648 session=1 capture_health attempt=1 audioMs=5013 samples=16043 rms=0.000000 peak=0.000000 silent=true inputUID=BuiltInMicrophoneDevice
[09:03:41.213] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[09:03:41.213] [INFO] [ASRBenchmark] ASR_BENCH t=2594483.866101 session=1 chunk_start index=7 ageMs=5509 samples=86016 newSamples=11434 audioMs=5376 provider=FluidAudio (Apple Silicon Optimized)
[09:03:41.213] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 86016) using FluidAudio (Apple Silicon Optimized)
[09:03:41.321] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=86016 audioMs=5376 elapsedMs=108 textChars=39 rtf=0.020
[09:03:41.321] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.11s
[09:03:41.321] [DEBUG] [ASRService] ✅ Streaming: 'One, two, three, four, five, six, seven' (0.11s)
[09:03:41.321] [INFO] [ASRBenchmark] ASR_BENCH t=2594483.974347 session=1 chunk_done index=7 elapsedMs=108 ageMs=5618 samples=86016 rawChars=39 cleanedChars=39 rtf=0.020
[09:03:41.321] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[09:03:41.846] [INFO] [ASRBenchmark] ASR_BENCH t=2594484.499230 session=1 capture_health attempt=1 audioMs=6016 samples=16042 rms=0.000000 peak=0.000000 silent=true inputUID=BuiltInMicrophoneDevice
[09:03:41.959] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[09:03:41.959] [INFO] [ASRBenchmark] ASR_BENCH t=2594484.612049 session=1 chunk_start index=8 ageMs=6255 samples=97963 newSamples=11947 audioMs=6123 provider=FluidAudio (Apple Silicon Optimized)
[09:03:41.959] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 97963) using FluidAudio (Apple Silicon Optimized)
[09:03:42.045] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=97963 audioMs=6123 elapsedMs=86 textChars=39 rtf=0.014
[09:03:42.045] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.09s
[09:03:42.045] [DEBUG] [ASRService] ✅ Streaming: 'One, two, three, four, five, six, seven' (0.09s)
[09:03:42.046] [INFO] [ASRBenchmark] ASR_BENCH t=2594484.698585 session=1 chunk_done index=8 elapsedMs=87 ageMs=6342 samples=97963 rawChars=39 cleanedChars=39 rtf=0.014
[09:03:42.046] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[09:03:42.684] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[09:03:42.684] [INFO] [ASRBenchmark] ASR_BENCH t=2594485.337253 session=1 chunk_start index=9 ageMs=6981 samples=109568 newSamples=11605 audioMs=6848 provider=FluidAudio (Apple Silicon Optimized)
[09:03:42.684] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 109568) using FluidAudio (Apple Silicon Optimized)
[09:03:42.738] [DEBUG] [ContentView] AI dictation prewarm complete slot=primary
[09:03:42.760] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=109568 audioMs=6848 elapsedMs=76 textChars=39 rtf=0.011
[09:03:42.760] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.08s
[09:03:42.761] [DEBUG] [ASRService] ✅ Streaming: 'One, two, three, four, five, six, seven' (0.08s)
[09:03:42.761] [INFO] [ASRBenchmark] ASR_BENCH t=2594485.413698 session=1 chunk_done index=9 elapsedMs=76 ageMs=7057 samples=109568 rawChars=39 cleanedChars=39 rtf=0.011
[09:03:42.761] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[09:03:42.849] [INFO] [ASRBenchmark] ASR_BENCH t=2594485.501906 session=1 capture_health attempt=1 audioMs=7019 samples=16043 rms=0.000000 peak=0.000000 silent=true inputUID=BuiltInMicrophoneDevice
[09:03:42.849] [INFO] [ASRBenchmark] ASR_BENCH t=2594485.501922 session=1 capture_health_recovery_triggered attempt=1 audioMs=7019 rms=0.000000 peak=0.000000
[09:03:42.849] [WARN] [ASRService] Audio route changed during capture; waiting for topology quiet (sustained silent PCM, generation=1, isStarting=false, isRunning=true)
[INFO] [DirectCoreAudioLifecycle] t=2594485.818219 Direct capture stop begin generation=1 reason=active_route_recovery retainPrepared=false
[09:03:43.167] [INFO] [ASRService] Recovering audio route after sustained silent PCM (generation=1)
[INFO] [DirectCoreAudioLifecycle] t=2594485.858366 Direct capture invalidate begin generation=1 reason=active_route_recovery
[DEBUG] [DirectCoreAudioLifecycle] t=2594485.858429 Direct capture listener drain generation=1 count=11 elapsedMs=0
[INFO] [DirectCoreAudioLifecycle] t=2594485.858608 Direct capture invalidate end generation=1 elapsedMs=0 reason=active_route_recovery
[INFO] [DirectCoreAudioLifecycle] t=2594485.858624 Direct capture stop end generation=1 elapsedMs=40 status=0 droppedPackets=0 retained=false
[09:03:43.206] [DEBUG] [ASRService] Audio engine retired (audio_route_recovery)
[INFO] [DirectCoreAudioLifecycle] t=2594485.880901 Direct capture device resolved off-main reason=recording_start device='MacBook Pro Microphone' id=110 elapsedMs=3
[INFO] [DirectCoreAudioLifecycle] t=2594485.881530 Direct capture prepare begin generation=2 device='MacBook Pro Microphone' reason=recording_start fingerprint={device=110 stream=111 virtual=48000Hz/1ch/32bit/flags=0x9/bytesPerFrame=4 physical=unavailable buffers=[1] nominalHz=48000 bufferFrames=512 variableMax=fixed dataSource=unavailable}
[DEBUG] [DirectCoreAudioLifecycle] t=2594485.912739 Direct capture listener unavailable generation=2 property=device_changed
[DEBUG] [DirectCoreAudioLifecycle] t=2594485.913047 Direct capture listener unavailable generation=2 property=variable_buffer_frame_size
[INFO] [DirectCoreAudioLifecycle] t=2594485.913705 Direct capture prepare end generation=2 elapsedMs=32 fingerprint={device=110 stream=111 virtual=48000Hz/1ch/32bit/flags=0x9/bytesPerFrame=4 physical=unavailable buffers=[1] nominalHz=48000 bufferFrames=512 variableMax=fixed dataSource=unavailable}
[INFO] [DirectCoreAudioLifecycle] t=2594485.916913 Direct capture start begin generation=2 device='MacBook Pro Microphone' validationMs=3
[INFO] [DirectCoreAudioLifecycle] t=2594485.972708 Direct capture start end generation=2 elapsedMs=56 fingerprint={device=110 stream=111 virtual=48000Hz/1ch/32bit/flags=0x9/bytesPerFrame=4 physical=unavailable buffers=[1] nominalHz=48000 bufferFrames=512 variableMax=fixed dataSource=unavailable}
[09:03:43.320] [INFO] [ASRBenchmark] ASR_BENCH t=2594485.972782 session=1 audio_backend kind=direct_core_audio device=110 generation=2 frames=512 sampleRate=48000 callbackMs=11
[09:03:43.328] [INFO] [ASRBenchmark] ASR_BENCH t=2594485.980980 session=1 attempt=2 first_audio sampleCount=171 frameLength=512 sampleRate=48000 bufferMs=11 acquisitionMs=98 elapsedMs=110
[09:03:43.328] [INFO] [ASRBenchmark] ASR_BENCH t=2594485.981037 session=1 route_recovery_first_pcm generation=1 attempt=1 attemptID=2 bufferedSamples=112470
[09:03:43.328] [INFO] [ASRService] Audio route recovery succeeded (generation=1, attempt=1)
[09:03:43.378] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[09:03:43.378] [INFO] [ASRBenchmark] ASR_BENCH t=2594486.030615 session=1 chunk_start index=10 ageMs=7674 samples=113153 newSamples=3585 audioMs=7072 provider=FluidAudio (Apple Silicon Optimized)
[09:03:43.378] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 113153) using FluidAudio (Apple Silicon Optimized)
[09:03:43.463] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=113153 audioMs=7072 elapsedMs=85 textChars=40 rtf=0.012
[09:03:43.463] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.09s
[09:03:43.464] [DEBUG] [ASRService] ✅ Streaming: 'One, two, three, four, five, six, seven.' (0.09s)
[09:03:43.464] [INFO] [ASRBenchmark] ASR_BENCH t=2594486.116615 session=1 chunk_done index=10 elapsedMs=86 ageMs=7760 samples=113153 rawChars=40 cleanedChars=40 rtf=0.012
[09:03:43.464] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[09:03:44.103] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[09:03:44.103] [INFO] [ASRBenchmark] ASR_BENCH t=2594486.756385 session=1 chunk_start index=11 ageMs=8400 samples=124758 newSamples=11605 audioMs=7797 provider=FluidAudio (Apple Silicon Optimized)
[09:03:44.103] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 124758) using FluidAudio (Apple Silicon Optimized)
[09:03:44.189] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=124758 audioMs=7797 elapsedMs=86 textChars=40 rtf=0.011
[09:03:44.189] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.09s
[09:03:44.190] [DEBUG] [ASRService] ✅ Streaming: 'One, two, three, four, five, six, seven.' (0.09s)
[09:03:44.190] [INFO] [ASRBenchmark] ASR_BENCH t=2594486.842761 session=1 chunk_done index=11 elapsedMs=86 ageMs=8486 samples=124758 rawChars=40 cleanedChars=40 rtf=0.011
[09:03:44.190] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[09:03:44.344] [INFO] [ASRBenchmark] ASR_BENCH t=2594486.997282 session=1 capture_health attempt=2 audioMs=1003 samples=16043 rms=0.000000 peak=0.000000 silent=true inputUID=BuiltInMicrophoneDevice
[09:03:44.795] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[09:03:44.795] [INFO] [ASRBenchmark] ASR_BENCH t=2594487.448507 session=1 chunk_start index=12 ageMs=9092 samples=135851 newSamples=11093 audioMs=8491 provider=FluidAudio (Apple Silicon Optimized)
[09:03:44.795] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 135851) using FluidAudio (Apple Silicon Optimized)
[09:03:44.882] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=135851 audioMs=8491 elapsedMs=86 textChars=40 rtf=0.010
[09:03:44.882] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.09s
[09:03:44.882] [DEBUG] [ASRService] ✅ Streaming: 'One, two, three, four, five, six, seven.' (0.09s)
[09:03:44.882] [INFO] [ASRBenchmark] ASR_BENCH t=2594487.535035 session=1 chunk_done index=12 elapsedMs=87 ageMs=9178 samples=135851 rawChars=40 cleanedChars=40 rtf=0.010
[09:03:44.882] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[09:03:45.323] [INFO] [ASRBenchmark] ASR_BENCH t=2594487.975896 session=1 capture_health attempt=2 audioMs=2005 samples=16043 rms=0.000000 peak=0.000000 silent=true inputUID=BuiltInMicrophoneDevice
[09:03:45.546] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[09:03:45.546] [INFO] [ASRBenchmark] ASR_BENCH t=2594488.199518 session=1 chunk_start index=13 ageMs=9843 samples=147798 newSamples=11947 audioMs=9237 provider=FluidAudio (Apple Silicon Optimized)
[09:03:45.546] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 147798) using FluidAudio (Apple Silicon Optimized)
[09:03:45.646] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=147798 audioMs=9237 elapsedMs=91 textChars=40 rtf=0.010
[09:03:45.646] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.10s
[09:03:45.647] [DEBUG] [ASRService] ✅ Streaming: 'One, two, three, four, five, six, seven.' (0.10s)
[09:03:45.647] [INFO] [ASRBenchmark] ASR_BENCH t=2594488.299616 session=1 chunk_done index=13 elapsedMs=100 ageMs=9943 samples=147798 rawChars=40 cleanedChars=40 rtf=0.011
[09:03:45.647] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[09:03:46.280] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[09:03:46.280] [INFO] [ASRBenchmark] ASR_BENCH t=2594488.932669 session=1 chunk_start index=14 ageMs=10576 samples=159574 newSamples=11776 audioMs=9973 provider=FluidAudio (Apple Silicon Optimized)
[09:03:46.280] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 159574) using FluidAudio (Apple Silicon Optimized)
[09:03:46.325] [INFO] [ASRBenchmark] ASR_BENCH t=2594488.978569 session=1 capture_health attempt=2 audioMs=3008 samples=16042 rms=0.000000 peak=0.000000 silent=true inputUID=BuiltInMicrophoneDevice
[09:03:46.367] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=159574 audioMs=9973 elapsedMs=87 textChars=40 rtf=0.009
[09:03:46.367] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.09s
[09:03:46.367] [DEBUG] [ASRService] ✅ Streaming: 'One, two, three, four, five, six, seven.' (0.09s)
[09:03:46.367] [INFO] [ASRBenchmark] ASR_BENCH t=2594489.019945 session=1 chunk_done index=14 elapsedMs=87 ageMs=10663 samples=159574 rawChars=40 cleanedChars=40 rtf=0.009
[09:03:46.367] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[09:03:47.007] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[09:03:47.007] [INFO] [ASRBenchmark] ASR_BENCH t=2594489.659898 session=1 chunk_start index=15 ageMs=11303 samples=171179 newSamples=11605 audioMs=10699 provider=FluidAudio (Apple Silicon Optimized)
[09:03:47.007] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 171179) using FluidAudio (Apple Silicon Optimized)
[09:03:47.092] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=171179 audioMs=10699 elapsedMs=85 textChars=40 rtf=0.008
[09:03:47.092] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.09s
[09:03:47.093] [DEBUG] [ASRService] ✅ Streaming: 'One, two, three, four, five, six, seven.' (0.09s)
[09:03:47.093] [INFO] [ASRBenchmark] ASR_BENCH t=2594489.745677 session=1 chunk_done index=15 elapsedMs=86 ageMs=11389 samples=171179 rawChars=40 cleanedChars=40 rtf=0.008
[09:03:47.093] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[09:03:47.328] [INFO] [ASRBenchmark] ASR_BENCH t=2594489.981318 session=1 capture_health attempt=2 audioMs=4011 samples=16043 rms=0.000000 peak=0.000000 silent=true inputUID=BuiltInMicrophoneDevice
[09:03:47.379] [DEBUG] [ContentView] stopAndProcessTranscription called
[09:03:47.379] [INFO] [ContentView] Output route selected: normal
[09:03:47.379] [INFO] [AppBenchmark] APP_BENCH t=2594490.032498 stop_path_enter route=normal
[09:03:47.385] [INFO] [ContentView] Routing decision snapshot | activeMode=dictate | rewrite=false | command=false | overlay=Dictation
[09:03:47.386] [DEBUG] [ContentView] Showing transcription processing state
[09:03:47.386] [INFO] [AppBenchmark] APP_BENCH t=2594490.038713 processing_ui_request status=Transcribing
[09:03:47.386] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594490.038727 manager set_processing_request processing=true overlayVisible=true active=false
[09:03:47.386] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594490.038861 notch text_update status=Transcribing
[09:03:47.386] [INFO] [AppBenchmark] APP_BENCH t=2594490.038889 processing_ui_requested status=Transcribing
[09:03:47.396] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594490.048913 manager processing_show_workitem_execute delayMs=0
[09:03:47.396] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594490.048955 notch set_processing processing=true state=idle bottomVisible=true commandExpanded=false
[09:03:47.396] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594490.048979 bottom_set_processing processing=true
[09:03:47.396] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594490.048992 notch set_processing_forwarded target=bottom
[09:03:47.396] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594490.048999 manager processing_show_workitem_return
[09:03:47.396] [INFO] [AppBenchmark] APP_BENCH t=2594490.049041 asr_stop_call
[09:03:47.396] [INFO] [ASRService] 🛑 STOP() called - beginning shutdown sequence
[09:03:47.396] [INFO] [ASRBenchmark] ASR_BENCH t=2594490.049054 session=1 stop_start ageMs=11692 bufferedSamples=177494
[09:03:47.396] [DEBUG] [ASRService] 📍 Preparing final transcription
[09:03:47.396] [DEBUG] [ASRService] 🚫 Setting isRunning = false...
[09:03:47.396] [DEBUG] [ASRService] ✅ isRunning disabled
[09:03:47.396] [DEBUG] [ASRService] 👁️ Stopping device monitoring...
[09:03:47.396] [DEBUG] [ASRService] ✅ Device monitoring stopped
[INFO] [DirectCoreAudioLifecycle] t=2594490.049110 Direct capture stop begin generation=2 reason=recording_stop retainPrepared=true
[INFO] [DirectCoreAudioLifecycle] t=2594490.067197 Direct capture stop end generation=2 elapsedMs=18 status=0 droppedPackets=0 retained=true
[09:03:47.478] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594490.131238 manager handle_state isRunning=false overlayVisible=true processing=true mode=Dictation
[09:03:47.478] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594490.131265 manager handle_state_return reason=processing_active
[09:03:47.478] [DEBUG] [ASRService] ♻️ Direct audio capture remains prepared
[09:03:47.562] [INFO] [AppBenchmark] APP_BENCH t=2594490.215096 sound_play_request sound=FV_end_0
[09:03:47.562] [INFO] [ASRBenchmark] ASR_BENCH t=2594490.215149 session=1 audio_capture_prepared retained=true phase=prepared generation=2
[09:03:47.562] [DEBUG] [ASRService] ⏳ Awaiting stopStreamingTimerAndAwait()...
[09:03:47.562] [INFO] [ASRBenchmark] ASR_BENCH t=2594490.215157 session=1 streaming_timer_stop begin
[09:03:47.562] [DEBUG] [ASRService] Streaming transcription task cancelled
[09:03:47.622] [INFO] [AppBenchmark] APP_BENCH t=2594490.275534 sound_play_dispatched sound=FV_end_0 elapsedMs=60
[09:03:47.630] [INFO] [ASRBenchmark] ASR_BENCH t=2594490.282941 session=1 streaming_timer_stop end elapsedMs=68 completedChunks=14
[09:03:47.630] [INFO] [ASRBenchmark] ASR_BENCH t=2594490.282973 session=1 stop_streaming_wait elapsedMs=68
[09:03:47.630] [DEBUG] [ASRService] ✅ stopStreamingTimerAndAwait() completed
[09:03:47.630] [INFO] [ASRBenchmark] ASR_BENCH t=2594490.282980 session=1 stop_audio_drained samples=177494 audioMs=11093
[09:03:47.630] [INFO] [ASRBenchmark] ASR_BENCH t=2594490.283006 session=1 silence_gate eligible=false skip=false reason=streaming_preview
[09:03:47.630] [INFO] [ASRBenchmark] ASR_BENCH t=2594490.283017 session=1 stop_ensure_ready skipped=true elapsedMs=0
[09:03:47.630] [DEBUG] [ASRService] Starting transcription with 177494 samples (11.093375 seconds)
[09:03:47.888] [INFO] [ASRBenchmark] ASR_BENCH provider_final_done samples=177494 audioMs=11093 elapsedMs=257 textChars=40 rtf=0.023 fallback=false source=full
[09:03:47.888] [DEBUG] [ASRService] stop(): final transcription finished source=full
[09:03:47.888] [DEBUG] [ASRService] Transcription completed: 'One, two, three, four, five, six, seven.' (confidence: 0.9691162)
[09:03:47.888] [INFO] [ASRService] Final ASR result | provider=FluidAudio (Apple Silicon Optimized) | samples=177494 | textChars=40 | confidence=0.9691162
[09:03:47.888] [INFO] [ASRBenchmark] ASR_BENCH t=2594490.540829 session=1 final_done elapsedMs=258 samples=177494 audioMs=11093 textChars=40 rtf=0.023 streamedChunks=14 source=full
[09:03:47.888] [DEBUG] [ASRService] After post-processing: 'One, two, three, four, five, six, seven.'
[09:03:47.888] [INFO] [ASRBenchmark] ASR_BENCH t=2594490.541019 session=1 stop_end result=success totalMs=492 recordingAgeMs=12184 cleanedChars=40
[09:03:47.888] [INFO] [AppBenchmark] APP_BENCH t=2594490.541085 asr_stop_return elapsedMs=492
[09:03:47.888] [INFO] [ContentView] Stop transcription result | chars=40 | empty=false
[09:03:47.888] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594490.541098 notch text_update status=empty
[09:03:47.898] [DEBUG] [ContentView] Routing transcription through AI post-processing
[09:03:47.898] [INFO] [AppBenchmark] APP_BENCH t=2594490.550668 processing_ui_request status=Refining
[09:03:47.898] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594490.550675 notch text_update status=Refining
[09:03:47.898] [INFO] [AppBenchmark] APP_BENCH t=2594490.550694 processing_ui_requested status=Refining
[09:03:48.010] [DEBUG] [ContentView] processTextWithAI using provider=fluid-1, model=fluid-1
[09:03:48.010] [DEBUG] [ContentView] [PromptTrace][Dictate] Private AI Provider task:
dictationEnhancement
[09:03:48.010] [DEBUG] [ContentView] [PromptTrace][Dictate] Input transcription (Q):
One, two, three, four, five, six, seven.
[09:03:48.010] [DEBUG] [ContentView] [PromptTrace][Dictate] Selected context text:
<none (dictation mode)>
[09:03:48.016] [DEBUG] [PrivateAIProvider] Private provider reused cached runtime model=fluid-1 backend=mlx:/Applications/FluidVoice_169_3.app/Contents/Helpers/fluid-intelligence-mlx:/Users/myuser/Library/Application Support/FluidIntelligence/Models/fluid-1-nvfp4-mlx:mtp:/Users/myuser/Library/Application Support/FluidIntelligence/Models/gemma-4-E2B-it-qat-assistant-bf16-mlx-mtp
[09:03:48.016] [INFO] [PrivateAIProvider] Private provider post-processing start backend=MLX model=fluid-1 inputChars=40 maxOutputTokens=256
[09:03:48.771] [INFO] [PrivateAIProvider] Private provider post-processing complete backend=MLX model=fluid-1 totalMs=753 ttftMs=326 promptTokens=1290 outputTokens=20 totalTokens=1310 tps=46.9
[09:03:48.771] [INFO] [PrivateAIProvider] Private provider post-processing detail backend=MLX mtpEnabled=1 mtpDraftTokens=20 mtpAcceptedTokens=16 mtpAcceptPercent=80 cacheHit=1 reusedTokens=0 prefixTokens=1271 prefillTokens=19 prefillMs=-1 clearMs=-1
[09:03:48.771] [DEBUG] [ContentView] [PromptTrace][Dictate] Model answer (A):
1, 2, 3, 4, 5, 6, 7.
[09:03:48.771] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594491.424082 notch text_update status=empty
[09:03:48.771] [INFO] [ContentView] Dictation AI post-processing finished in 874ms provider=fluid-1 model=fluid-1 inputChars=40 fallback=false
[09:03:48.771] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594491.424209 notch text_update status=empty
[09:03:48.771] [INFO] [ContentView] Transcription finalized (chars: 21)
[09:03:48.771] [INFO] [AppBenchmark] APP_BENCH t=2594491.424356 transcription_finalized chars=21
[09:03:48.771] [INFO] [AppBenchmark] APP_BENCH t=2594491.424367 text_ready chars=21
[09:03:48.798] [DEBUG] [TranscriptionHistoryStore] Added transcription to history (total: 2765)
[09:03:48.798] [DEBUG] [ContentView] Typing decision → frontmost: FluidVoice, fluidFrontmost: true, editorFocused: true, willTypeExternally: false
[09:03:48.798] [INFO] [AppBenchmark] APP_BENCH t=2594491.451384 overlay_hide_request reason=after_output lifecycle=1
[09:03:48.916] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594491.569488 notch set_processing processing=false state=idle bottomVisible=true commandExpanded=false
[09:03:48.916] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594491.569545 bottom_set_processing processing=false
[09:03:48.916] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594491.569555 notch set_processing_forwarded target=bottom
[09:03:48.916] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594491.569559 manager finish_hide_request
[09:03:48.917] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594491.569582 notch hide_called state=idle bottomVisible=true
[09:03:48.917] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594491.569645 bottom_hide_start windowExists=true
[09:03:48.917] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594491.569666 bottom_hide_animation_start
[09:03:48.952] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594491.605103 bottom_hide_complete elapsedMs=35
[09:03:48.952] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594491.605139 notch hide_return reason=not_visible state=idle notchExists=false
[09:03:48.952] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594491.605146 notch hide_complete target=none elapsedMs=36
[09:03:48.952] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594491.605233 manager finish_hide_complete outcome=hidden elapsedMs=36
[09:03:48.952] [INFO] [AppBenchmark] APP_BENCH t=2594491.605242 overlay_hidden reason=after_output elapsedMs=36
[09:03:56.444] [DEBUG] [ContentView] Mode transition: Optional(FluidVoice.SidebarItem.welcome) → Optional(FluidVoice.SidebarItem.changelog)
[09:03:56.444] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594499.096792 manager set_mode mode=Dictation
[09:03:56.444] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594499.096836 notch set_mode mode=Dictation
[09:03:59.773] [DEBUG] [ContentView] Mode transition: Optional(FluidVoice.SidebarItem.changelog) → Optional(FluidVoice.SidebarItem.preferences)
[09:03:59.773] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594502.425836 manager set_mode mode=Dictation
[09:03:59.773] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2594502.425889 notch set_mode mode=Dictation
[09:04:31.064] [DEBUG] [AppDelegate] Automatic update check failed: The request timed out.

---

UPD: the voice capture DOES work properly with my bluetooth headphones, it doesn't stop after 3-4 seconds and seems to work properly. Switching back to MacBook's standard microphone returns the issue. So it looks like this issue is related to the built-in mic.

[altic-dev]

Hey @Defake, thanks for the logs. It was really helpful. Can you try an external mic (wired or bluetooth) and report back if you run into the same issue ? If you do run into the same issue, can share additional logs.

Our current hypothesis is that there is an IOProc failure with the built-in mac audio input which we don't have much control over.

[Defake]

Hello @altic-dev , I did report in my previous comment about the external mic. Not sure why this paragraph was not visible, this is it:

the voice capture DOES work properly with my bluetooth headphones, it doesn't stop after 3-4 seconds and seems to work properly. Switching back to MacBook's standard microphone returns the issue. So it looks like this issue is related to the built-in mic.

[altic-dev]

Yeah, not sure if we can help in this case unfortunately. The logs show there's nothing wrong with app itself and more on sound stopping from your mic for some reason

[Defake]

@altic-dev but it's working correctly in 1.6.6. The 1.6.7 brought this bug, so the issue certainly was introduced by some commits in between. I thought it should help finding it.

[grohith327]

@altic-dev can you share a build to @Defake with the changes from branch https://github.com/altic-dev/FluidVoice/tree/852-fix ??

[altic-dev]

@Defake here is the link to the dmg - https://drive.google.com/file/d/1-I31w2wqOlRC1WjDKqDIfp95YMwU8IJY/view?usp=share_link

Please try it out and let us know if you run into the same issue while using your mac mic

[Defake]

@altic-dev , @grohith327 thank you, I tried the new version, unfortunately the behavior is still the same (dictated from 1 to 15, it stopped hearing me after 8), here are the logs:

Details

[RUN] PID=84995 AppVersion=1.6.9 Build=20 OS=Version 26.5.2 (Build 25F84)
[RUN] Processed by com.FluidApp.app
[11:07:20.900] [INFO] [AppServices] 📦 AppServices singleton created (services not yet initialized)
[11:07:21.244] [INFO] [AppServices] 🎤 Lazily creating ASRService
[11:07:21.246] [INFO] [AppServices] 🔊 Lazily creating AudioHardwareObserver
[11:07:21.433] [DEBUG] [GlobalHotkeyManager] Starting delayed initialization...
[11:07:21.433] [INFO] [GlobalHotkeyManager] Hotkey activation mode set to Toggle
[11:07:21.479] [DEBUG] [ContentView] Accessibility enabled, reinitializing hotkey manager
[11:07:21.479] [INFO] [GlobalHotkeyManager] Manual reinitialization requested
[11:07:21.479] [DEBUG] [GlobalHotkeyManager] Starting delayed initialization...
[11:07:21.479] [DEBUG] [ContentView] Mode transition: nil → Optional(FluidVoice.SidebarItem.welcome)
[11:07:21.479] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848372.909925 manager set_mode mode=Dictation
[11:07:21.479] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848372.909961 notch set_mode mode=Dictation
[11:07:21.714] [INFO] [AppDelegate] Application launched [loginItemLaunch=false]
[11:07:21.803] [INFO] [SettingsStore] Initializing app with dock visibility: true
[11:07:21.803] [INFO] [LocalAPIServer] Local API disabled
[11:07:21.803] [INFO] [AppDelegate] Scheduling automatic update check...
[11:07:21.967] [DEBUG] [AppDelegate] Brought main window to front
[11:07:21.970] [DEBUG] [GlobalHotkeyManager] Setup attempt 1/5
[11:07:22.024] [INFO] [GlobalHotkeyManager] Event tap successfully created and enabled
[11:07:22.024] [INFO] [GlobalHotkeyManager] Successfully initialized on attempt 1
[11:07:22.940] [INFO] [ContentView] 🚦 Startup delay complete, signaling UI ready...
[11:07:22.940] [INFO] [AppServices] 🚦 UI Ready signal received - services can now initialize
[11:07:23.024] [DEBUG] [ContentView] Skipping auto-preload - models downloaded manually via UI
[11:07:23.479] [DEBUG] [GlobalHotkeyManager] Setup attempt 1/5
[11:07:23.482] [INFO] [GlobalHotkeyManager] Event tap successfully created and enabled
[11:07:23.482] [INFO] [GlobalHotkeyManager] Successfully initialized on attempt 1
[11:07:24.154] [DEBUG] [ASRService] Models exist on disk: true
[11:07:24.218] [INFO] [ContentView] 🔊 Starting gated audio initialization...
[11:07:24.345] [INFO] [AudioHardwareObserver] Clamshell state listener registered (closed=false)
[11:07:24.436] [DEBUG] [ASRService] Device list change listener registered
[INFO] [DirectCoreAudioLifecycle] t=2848375.879697 Direct capture device resolved off-main reason=prepare:startup device='MacBook Pro Microphone' id=110 elapsedMs=4
[INFO] [DirectCoreAudioLifecycle] t=2848375.880301 Direct capture prepare begin generation=1 device='MacBook Pro Microphone' transport=device_ioproc reason=startup fingerprint={device=110 stream=111 virtual=48000Hz/1ch/32bit/flags=0x9/bytesPerFrame=4 physical=unavailable buffers=[1] nominalHz=48000 bufferFrames=512 variableMax=fixed dataSource=unavailable}
[DEBUG] [DirectCoreAudioLifecycle] t=2848375.896703 Direct capture listener unavailable generation=1 property=device_changed
[DEBUG] [DirectCoreAudioLifecycle] t=2848375.897028 Direct capture listener unavailable generation=1 property=variable_buffer_frame_size
[INFO] [DirectCoreAudioLifecycle] t=2848375.897802 Direct capture prepare end generation=1 transport=device_ioproc elapsedMs=18 fingerprint={device=110 stream=111 virtual=48000Hz/1ch/32bit/flags=0x9/bytesPerFrame=4 physical=unavailable buffers=[1] nominalHz=48000 bufferFrames=512 variableMax=fixed dataSource=unavailable}
[11:07:24.481] [INFO] [ASRService] Prepared direct Core Audio input 'MacBook Pro Microphone' (48000Hz, 512 frames, generation=1, transport=device_ioproc, reason=startup)
[11:07:24.481] [INFO] [ASRBenchmark] ASR_BENCH t=2848375.911305 session=0 direct_audio_prewarm reason=startup elapsedMs=45
[11:07:24.481] [INFO] [ContentView] ✅ Audio subsystems initialized
[11:07:24.482] [DEBUG] [ASRService] Models exist on disk: true
[11:07:24.482] [INFO] [ASRService] Models found on disk, auto-loading...
[11:07:24.482] [INFO] [ASRService] ASRService: Created FluidAudio provider [vocabBoosting=false]
[11:07:24.482] [INFO] [ASRService] ensureAsrReady() requested for model=parakeet-tdt [supportsStreaming=true] provider=parakeet-tdt:FluidAudioProvider:FluidAudio (Apple Silicon Optimized)
[11:07:24.578] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848376.008717 bottom_prepared
[11:07:24.579] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848376.009131 manager handle_state isRunning=false overlayVisible=false processing=false mode=Dictation
[11:07:24.579] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848376.009150 manager handle_state_return reason=visibility_unchanged
[11:07:24.579] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848376.009163 notch text_update status=empty
[11:07:24.619] [DEBUG] [ASRService] ensureAsrReady(begin): provider=FluidAudio (Apple Silicon Optimized), providerReady=false, isAsrReady=false, isRunning=false
[11:07:24.621] [INFO] [ASRService] === ASR INITIALIZATION START ===
[11:07:24.621] [INFO] [ASRService] Using provider: FluidAudio (Apple Silicon Optimized) [providerReady=false]
[11:07:24.621] [INFO] [ASRService] Models already cached on disk: true
[11:07:24.621] [DEBUG] [ASRService] Model cache lookup complete in 0.000s
[11:07:24.621] [INFO] [ASRService] 📦 LOADING cached model into memory...
[11:07:24.621] [INFO] [ASRService] Calling transcriptionProvider.prepare()...
[11:07:24.621] [INFO] [FluidAudioProvider] FluidAudioProvider: Starting model preparation for Parakeet TDT v3 (Multilingual) [version=v3]
[11:07:24.621] [DEBUG] [FluidAudioProvider] FluidAudioProvider: target cache directory=/Users/myuser/Library/Application Support/FluidAudio/Models
[11:07:24.969] [DEBUG] [FluidAudioProvider] FluidAudioProvider: Models downloadAndLoad returned in 0.35s
[11:07:25.030] [DEBUG] [FluidAudioProvider] FluidAudioProvider: Streaming AsrManager initialized
[11:07:25.030] [DEBUG] [FluidAudioProvider] FluidAudioProvider: Word boosting disabled by configuration
[11:07:25.030] [INFO] [FluidAudioProvider] FluidAudioProvider: Models ready [isWordBoostingActive=false, terms=0]
[11:07:25.030] [INFO] [ASRService] ASRService: Provider 'FluidAudio (Apple Silicon Optimized)' prepared successfully in 0.41s
[11:07:25.030] [INFO] [ASRService] ✓ Provider preparation completed in 0.4 seconds
[11:07:25.030] [INFO] [ASRService] ⏳ Model loaded, waiting for first transcription to complete...
[11:07:25.030] [INFO] [ASRService] === ASR INITIALIZATION COMPLETE ===
[11:07:25.030] [INFO] [ASRService] Total initialization time: 0.4 seconds
[11:07:25.030] [DEBUG] [ContentView] Initial hotkey manager health check: true
[11:07:25.030] [INFO] [AppDelegate] Performing automatic update check for altic-dev/Fluid-oss
[11:07:25.097] [INFO] [ASRService] Models auto-loaded successfully on startup
[11:07:25.097] [DEBUG] [ASRService] Audio capture prewarm skipped - backend already prepared
[11:07:31.378] [INFO] [ContentView] ContentView: startRecording() for model=Parakeet TDT v3 (Multilingual), supportsStreaming=true
[11:07:31.380] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848382.807671 manager set_mode mode=Dictation
[11:07:31.380] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848382.807779 notch set_mode mode=Dictation
[11:07:31.380] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848382.808024 manager instant_show_request mode=Dictation
[11:07:31.380] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848382.808049 manager show_workitem_execute mode=Dictation
[11:07:31.380] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848382.808571 notch show_called mode=Dictation state=idle commandExpanded=false
[11:07:31.380] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848382.808589 notch show_internal_enter mode=Dictation state=idle
[11:07:31.380] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848382.809071 notch show_internal_route target=bottom
[11:07:31.380] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848382.809088 notch bottom_route_start mode=Dictation
[11:07:31.380] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848382.809098 bottom_show_start mode=Dictation windowExists=true
[11:07:31.388] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848382.818589 bottom_order_front elapsedMs=9
[11:07:31.388] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848382.818610 bottom_visible elapsedMs=10
[11:07:31.388] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848382.818662 notch bottom_route_return elapsedMs=10
[11:07:31.388] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848382.818667 manager show_workitem_return mode=Dictation
[11:07:31.388] [INFO] [AppBenchmark] APP_BENCH t=2848382.818670 overlay_phase phase=connecting
[11:07:31.389] [INFO] [ASRService] 🎤 START() called - beginning recording session
[11:07:31.390] [INFO] [ASRBenchmark] ASR_BENCH t=2848382.820177 session=0 route_recovery_handoff_end generations=[] elapsedMs=0 capturePhase=prepared
[11:07:31.390] [DEBUG] [ASRService] 🧹 Clearing buffers and state
[11:07:31.390] [INFO] [ASRBenchmark] ASR_BENCH t=2848382.820249 session=1 recording_start model=parakeet-tdt provider=nvidia supportsStreaming=true
[11:07:31.390] [DEBUG] [ASRService] ✅ Buffers cleared
[11:07:31.394] [DEBUG] [ContentView] ContentView: pre-load model task started
[11:07:31.394] [INFO] [ASRService] ensureAsrReady() requested for model=parakeet-tdt [supportsStreaming=true] provider=parakeet-tdt:FluidAudioProvider:FluidAudio (Apple Silicon Optimized)
[11:07:31.458] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848382.888733 notch text_update status=empty
[11:07:31.458] [DEBUG] [ASRService] ensureAsrReady(begin): provider=FluidAudio (Apple Silicon Optimized), providerReady=true, isAsrReady=true, isRunning=false
[11:07:31.458] [DEBUG] [ASRService] ASR already ready with loaded models, skipping initialization
[INFO] [DirectCoreAudioLifecycle] t=2848382.892878 Direct capture device resolved off-main reason=recording_start device='MacBook Pro Microphone' id=110 elapsedMs=4
[11:07:31.522] [DEBUG] [ContentView] Model pre-loaded during recording
[DEBUG] [DirectCoreAudioLifecycle] t=2848382.953242 Direct capture prepare reuse generation=1 transport=device_ioproc reason=recording_start fingerprint={device=110 stream=111 virtual=48000Hz/1ch/32bit/flags=0x9/bytesPerFrame=4 physical=unavailable buffers=[1] nominalHz=48000 bufferFrames=512 variableMax=fixed dataSource=unavailable}
[INFO] [DirectCoreAudioLifecycle] t=2848382.953605 Direct capture start begin generation=1 device='MacBook Pro Microphone' validationMs=0
[INFO] [DirectCoreAudioLifecycle] t=2848383.013242 Direct capture start end generation=1 elapsedMs=60 fingerprint={device=110 stream=111 virtual=48000Hz/1ch/32bit/flags=0x9/bytesPerFrame=4 physical=unavailable buffers=[1] nominalHz=48000 bufferFrames=512 variableMax=fixed dataSource=unavailable}
[11:07:31.614] [INFO] [ASRBenchmark] ASR_BENCH t=2848383.044147 session=1 audio_backend kind=direct_core_audio device=110 transport=device_ioproc generation=1 frames=512 sampleRate=48000 callbackMs=11
[11:07:31.614] [INFO] [ASRBenchmark] ASR_BENCH t=2848383.044173 session=1 first_pcm_wait_begin attempt=1 attemptID=1 timeoutMs=2000 routeGeneration=0 captureGeneration=1
[11:07:31.614] [INFO] [ASRBenchmark] ASR_BENCH t=2848383.044205 session=1 attempt=1 first_audio sampleCount=171 frameLength=512 sampleRate=48000 bufferMs=11 acquisitionMs=188 elapsedMs=200
[11:07:31.614] [DEBUG] [MicrophonePreferenceCoordinator] Confirmed active microphone as 'MacBook Pro Microphone'
[11:07:31.614] [INFO] [ASRBenchmark] ASR_BENCH t=2848383.044304 session=1 first_pcm_wait_end result=ready attempt=1 attemptID=1 bufferedSamples=512
[11:07:31.614] [INFO] [ASRService] ✅ Audio capture running after first PCM (session=1)
[11:07:31.614] [INFO] [AppBenchmark] APP_BENCH t=2848383.044389 sound_play_request sound=FV_start_0
[11:07:31.665] [DEBUG] [ContentView] Captured recording app context: app=FluidVoice, bundleId=com.FluidApp.app, title=FluidVoice
[11:07:31.688] [DEBUG] [ContentView] Captured preceding text for continuous dictation (chars=0)
[11:07:31.695] [INFO] [AppBenchmark] APP_BENCH t=2848383.125629 sound_play_dispatched sound=FV_start_0 elapsedMs=81
[11:07:31.695] [INFO] [AppBenchmark] APP_BENCH t=2848383.125925 overlay_phase phase=recording trigger=first_pcm
[11:07:31.808] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848383.238771 manager handle_state isRunning=true overlayVisible=true processing=false mode=Dictation
[11:07:31.808] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848383.238796 manager handle_state_return reason=visibility_unchanged
[11:07:31.808] [DEBUG] [ContentView] ContentView: AI dictation prewarm started slot=primary
[11:07:31.809] [DEBUG] [MediaPlaybackService] MediaPlaybackService: Track info received
- App: Arc
- Bundle: company.thebrowser.Browser
- Title: A movie
- isPlaying: false
- playbackRate: 0.0
- Determined playing: false
- elapsedMs: 112
[11:07:31.809] [DEBUG] [MediaPlaybackService] MediaPlaybackService: Media is not playing, no action needed (elapsedMs: 112)
[11:07:31.840] [DEBUG] [ASRService] 📡 Starting streaming transcription...
[11:07:31.840] [INFO] [ASRBenchmark] ASR_BENCH t=2848383.270162 session=1 streaming_timer_start intervalMs=600 minSamples=16000
[11:07:31.840] [DEBUG] [ASRService] Starting streaming transcription task (interval: 0.6s, minSamples: 16000)
[11:07:31.840] [INFO] [ASRService] ✅ START() completed successfully
[11:07:32.209] [DEBUG] [ASRService] 🔄 runStreamingLoop() - ENTERED
[11:07:32.209] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[11:07:32.209] [DEBUG] [ASRService] Waiting for more audio data (10070/16000 samples)
[11:07:32.209] [INFO] [ASRBenchmark] ASR_BENCH t=2848383.639485 session=1 chunk_wait index=1 ageMs=819 samples=10070 minSamples=16000
[11:07:32.209] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[11:07:32.280] [INFO] [PrivateAIProvider] Private provider creating runtime model=fluid-1 backend=mlx:/Applications/FluidVoice_169_aug17.app/Contents/Helpers/fluid-intelligence-mlx:/Users/myuser/Library/Application Support/FluidIntelligence/Models/fluid-1-nvfp4-mlx:mtp:/Users/myuser/Library/Application Support/FluidIntelligence/Models/gemma-4-E2B-it-qat-assistant-bf16-mlx-mtp previous=none
[11:07:32.582] [INFO] [ASRBenchmark] ASR_BENCH t=2848384.012586 session=1 capture_health attempt=1 audioMs=1003 samples=16043 rms=0.049009 peak=0.248678 silent=false inputUID=BuiltInMicrophoneDevice
[11:07:32.848] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[11:07:32.848] [INFO] [ASRBenchmark] ASR_BENCH t=2848384.278965 session=1 chunk_start index=2 ageMs=1459 samples=20139 newSamples=20139 audioMs=1259 provider=FluidAudio (Apple Silicon Optimized)
[11:07:32.848] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 20139) using FluidAudio (Apple Silicon Optimized)
[11:07:33.239] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=20139 audioMs=1259 elapsedMs=390 textChars=0 rtf=0.310
[11:07:33.239] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.39s
[11:07:33.239] [INFO] [ASRBenchmark] ASR_BENCH t=2848384.669868 session=1 chunk_done index=2 elapsedMs=391 ageMs=1850 samples=20139 rawChars=0 cleanedChars=0 rtf=0.310
[11:07:33.239] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[11:07:33.239] [INFO] [ASRService] ✅ Model warmed up - first streaming transcription completed
[11:07:33.585] [INFO] [ASRBenchmark] ASR_BENCH t=2848385.015255 session=1 capture_health attempt=1 audioMs=2005 samples=16043 rms=0.035040 peak=0.142259 silent=false inputUID=BuiltInMicrophoneDevice
[11:07:33.860] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[11:07:33.861] [INFO] [ASRBenchmark] ASR_BENCH t=2848385.291036 session=1 chunk_start index=3 ageMs=2471 samples=36352 newSamples=16213 audioMs=2272 provider=FluidAudio (Apple Silicon Optimized)
[11:07:33.861] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 36352) using FluidAudio (Apple Silicon Optimized)
[11:07:34.029] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=36352 audioMs=2272 elapsedMs=169 textChars=31 rtf=0.074
[11:07:34.030] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.17s
[11:07:34.030] [DEBUG] [ASRService] ✅ Streaming: 'Oh one, two, three, four, five.' (0.17s)
[11:07:34.041] [INFO] [ASRBenchmark] ASR_BENCH t=2848385.460152 session=1 chunk_done index=3 elapsedMs=169 ageMs=2640 samples=36352 rawChars=31 cleanedChars=31 rtf=0.074
[11:07:34.041] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[11:07:34.587] [INFO] [ASRBenchmark] ASR_BENCH t=2848386.017846 session=1 capture_health attempt=1 audioMs=3008 samples=16042 rms=0.055403 peak=0.197646 silent=false inputUID=BuiltInMicrophoneDevice
[11:07:34.670] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[11:07:34.670] [INFO] [ASRBenchmark] ASR_BENCH t=2848386.100305 session=1 chunk_start index=4 ageMs=3280 samples=49323 newSamples=12971 audioMs=3083 provider=FluidAudio (Apple Silicon Optimized)
[11:07:34.670] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 49323) using FluidAudio (Apple Silicon Optimized)
[11:07:34.742] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=49323 audioMs=3083 elapsedMs=72 textChars=35 rtf=0.023
[11:07:34.742] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.07s
[11:07:34.742] [DEBUG] [ASRService] ✅ Streaming: 'A one, two, three, four, five, six.' (0.07s)
[11:07:34.742] [INFO] [ASRBenchmark] ASR_BENCH t=2848386.172938 session=1 chunk_done index=4 elapsedMs=73 ageMs=3353 samples=49323 rawChars=35 cleanedChars=35 rtf=0.023
[11:07:34.742] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[11:07:35.351] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[11:07:35.351] [INFO] [ASRBenchmark] ASR_BENCH t=2848386.781863 session=1 chunk_start index=5 ageMs=3962 samples=60246 newSamples=10923 audioMs=3765 provider=FluidAudio (Apple Silicon Optimized)
[11:07:35.351] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 60246) using FluidAudio (Apple Silicon Optimized)
[11:07:35.469] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=60246 audioMs=3765 elapsedMs=112 textChars=47 rtf=0.030
[11:07:35.469] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.12s
[11:07:35.469] [DEBUG] [ASRService] ✅ Streaming: 'One, two, three, four, five, six, seven, eight.' (0.12s)
[11:07:35.469] [INFO] [ASRBenchmark] ASR_BENCH t=2848386.899764 session=1 chunk_done index=5 elapsedMs=118 ageMs=4079 samples=60246 rawChars=47 cleanedChars=47 rtf=0.031
[11:07:35.469] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[11:07:35.590] [INFO] [ASRBenchmark] ASR_BENCH t=2848387.020587 session=1 capture_health attempt=1 audioMs=4011 samples=16043 rms=0.053456 peak=0.190141 silent=false inputUID=BuiltInMicrophoneDevice
[11:07:36.108] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[11:07:36.108] [INFO] [ASRBenchmark] ASR_BENCH t=2848387.538281 session=1 chunk_start index=6 ageMs=4718 samples=72363 newSamples=12117 audioMs=4523 provider=FluidAudio (Apple Silicon Optimized)
[11:07:36.108] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 72363) using FluidAudio (Apple Silicon Optimized)
[11:07:36.183] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=72363 audioMs=4523 elapsedMs=75 textChars=48 rtf=0.017
[11:07:36.183] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.08s
[11:07:36.183] [DEBUG] [ASRService] ✅ Streaming: 'A one, two, three, four, five, six, seven, eight' (0.08s)
[11:07:36.183] [INFO] [ASRBenchmark] ASR_BENCH t=2848387.613520 session=1 chunk_done index=6 elapsedMs=75 ageMs=4793 samples=72363 rawChars=48 cleanedChars=48 rtf=0.017
[11:07:36.183] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[11:07:36.593] [INFO] [ASRBenchmark] ASR_BENCH t=2848388.023233 session=1 capture_health attempt=1 audioMs=5013 samples=16043 rms=0.000000 peak=0.000000 silent=true inputUID=BuiltInMicrophoneDevice
[11:07:36.822] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[11:07:36.822] [INFO] [ASRBenchmark] ASR_BENCH t=2848388.252492 session=1 chunk_start index=7 ageMs=5432 samples=83798 newSamples=11435 audioMs=5237 provider=FluidAudio (Apple Silicon Optimized)
[11:07:36.822] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 83798) using FluidAudio (Apple Silicon Optimized)
[11:07:36.921] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=83798 audioMs=5237 elapsedMs=99 textChars=46 rtf=0.019
[11:07:36.921] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.10s
[11:07:36.921] [DEBUG] [ASRService] ✅ Streaming: 'One, two, three, four, five, six, seven, eight' (0.10s)
[11:07:36.922] [INFO] [ASRBenchmark] ASR_BENCH t=2848388.352138 session=1 chunk_done index=7 elapsedMs=100 ageMs=5532 samples=83798 rawChars=46 cleanedChars=46 rtf=0.019
[11:07:36.922] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[11:07:37.560] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[11:07:37.560] [INFO] [ASRBenchmark] ASR_BENCH t=2848388.990565 session=1 chunk_start index=8 ageMs=6170 samples=95574 newSamples=11776 audioMs=5973 provider=FluidAudio (Apple Silicon Optimized)
[11:07:37.560] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 95574) using FluidAudio (Apple Silicon Optimized)
[11:07:37.595] [INFO] [ASRBenchmark] ASR_BENCH t=2848389.025982 session=1 capture_health attempt=1 audioMs=6016 samples=16042 rms=0.000000 peak=0.000000 silent=true inputUID=BuiltInMicrophoneDevice
[11:07:37.636] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=95574 audioMs=5973 elapsedMs=76 textChars=47 rtf=0.013
[11:07:37.636] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.08s
[11:07:37.636] [DEBUG] [ASRService] ✅ Streaming: 'One, two, three, four, five, six, seven, eight.' (0.08s)
[11:07:37.636] [INFO] [ASRBenchmark] ASR_BENCH t=2848389.066806 session=1 chunk_done index=8 elapsedMs=76 ageMs=6247 samples=95574 rawChars=47 cleanedChars=47 rtf=0.013
[11:07:37.636] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[11:07:37.782] [DEBUG] [ContentView] AI dictation prewarm complete slot=primary
[11:07:38.272] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[11:07:38.272] [INFO] [ASRBenchmark] ASR_BENCH t=2848389.702863 session=1 chunk_start index=9 ageMs=6883 samples=107008 newSamples=11434 audioMs=6688 provider=FluidAudio (Apple Silicon Optimized)
[11:07:38.272] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 107008) using FluidAudio (Apple Silicon Optimized)
[11:07:38.357] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=107008 audioMs=6688 elapsedMs=85 textChars=46 rtf=0.013
[11:07:38.358] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.09s
[11:07:38.358] [DEBUG] [ASRService] ✅ Streaming: 'One, two, three, four, five, six, seven, eight' (0.09s)
[11:07:38.358] [INFO] [ASRBenchmark] ASR_BENCH t=2848389.788412 session=1 chunk_done index=9 elapsedMs=86 ageMs=6968 samples=107008 rawChars=46 cleanedChars=46 rtf=0.013
[11:07:38.358] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[11:07:38.598] [INFO] [ASRBenchmark] ASR_BENCH t=2848390.028472 session=1 capture_health attempt=1 audioMs=7019 samples=16043 rms=0.000000 peak=0.000000 silent=true inputUID=BuiltInMicrophoneDevice
[11:07:38.598] [INFO] [ASRBenchmark] ASR_BENCH t=2848390.028488 session=1 capture_health_recovery_triggered attempt=1 audioMs=7019 rms=0.000000 peak=0.000000 inputUID=BuiltInMicrophoneDevice from=device_ioproc to=auhal
[11:07:38.598] [WARN] [ASRService] Audio route changed during capture; waiting for topology quiet (sustained silent PCM, generation=1, isStarting=false, isRunning=true)
[INFO] [DirectCoreAudioLifecycle] t=2848390.346640 Direct capture stop begin generation=1 reason=active_route_recovery retainPrepared=false
[11:07:38.916] [INFO] [ASRService] Recovering audio route after sustained silent PCM (generation=1)
[INFO] [DirectCoreAudioLifecycle] t=2848390.362855 Direct capture invalidate begin generation=1 reason=active_route_recovery
[DEBUG] [DirectCoreAudioLifecycle] t=2848390.362902 Direct capture listener drain generation=1 count=11 elapsedMs=0
[INFO] [DirectCoreAudioLifecycle] t=2848390.364089 Direct capture invalidate end generation=1 elapsedMs=1 reason=active_route_recovery
[INFO] [DirectCoreAudioLifecycle] t=2848390.364128 Direct capture stop end generation=1 elapsedMs=17 status=0 droppedPackets=0 retained=false
[11:07:38.933] [DEBUG] [ASRService] Audio engine retired (audio_route_recovery)
[INFO] [DirectCoreAudioLifecycle] t=2848390.377091 Direct capture device resolved off-main reason=recording_start device='MacBook Pro Microphone' id=110 elapsedMs=3
[INFO] [DirectCoreAudioLifecycle] t=2848390.377528 Direct capture prepare begin generation=2 device='MacBook Pro Microphone' transport=auhal reason=recording_start fingerprint={device=110 stream=111 virtual=48000Hz/1ch/32bit/flags=0x9/bytesPerFrame=4 physical=unavailable buffers=[1] nominalHz=48000 bufferFrames=512 variableMax=fixed dataSource=unavailable}
[11:07:38.991] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[11:07:38.991] [INFO] [ASRBenchmark] ASR_BENCH t=2848390.421692 session=1 chunk_start index=10 ageMs=7601 samples=112299 newSamples=5291 audioMs=7019 provider=FluidAudio (Apple Silicon Optimized)
[11:07:38.991] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 112299) using FluidAudio (Apple Silicon Optimized)
[DEBUG] [DirectCoreAudioLifecycle] t=2848390.436284 Direct capture listener unavailable generation=2 property=device_changed
[DEBUG] [DirectCoreAudioLifecycle] t=2848390.436548 Direct capture listener unavailable generation=2 property=variable_buffer_frame_size
[INFO] [DirectCoreAudioLifecycle] t=2848390.437122 Direct capture prepare end generation=2 transport=auhal elapsedMs=60 fingerprint={device=110 stream=111 virtual=48000Hz/1ch/32bit/flags=0x9/bytesPerFrame=4 physical=unavailable buffers=[1] nominalHz=48000 bufferFrames=512 variableMax=fixed dataSource=unavailable}
[INFO] [DirectCoreAudioLifecycle] t=2848390.437565 Direct capture start begin generation=2 device='MacBook Pro Microphone' validationMs=0
[INFO] [DirectCoreAudioLifecycle] t=2848390.485343 Direct capture start end generation=2 elapsedMs=48 fingerprint={device=110 stream=111 virtual=48000Hz/1ch/32bit/flags=0x9/bytesPerFrame=4 physical=unavailable buffers=[1] nominalHz=48000 bufferFrames=512 variableMax=fixed dataSource=unavailable}
[11:07:39.097] [INFO] [ASRBenchmark] ASR_BENCH t=2848390.527359 session=1 audio_backend kind=direct_core_audio device=110 transport=auhal generation=2 frames=512 sampleRate=48000 callbackMs=11
[11:07:39.098] [INFO] [ASRBenchmark] ASR_BENCH t=2848390.527413 session=1 attempt=2 first_audio sampleCount=171 frameLength=512 sampleRate=48000 bufferMs=11 acquisitionMs=109 elapsedMs=122
[11:07:39.098] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=112299 audioMs=7019 elapsedMs=106 textChars=47 rtf=0.015
[11:07:39.103] [INFO] [ASRBenchmark] ASR_BENCH t=2848390.533201 session=1 route_recovery_first_pcm generation=1 attempt=1 attemptID=2 bufferedSamples=112982
[11:07:39.103] [INFO] [ASRService] Audio route recovery succeeded (generation=1, attempt=1)
[11:07:39.103] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.11s
[11:07:39.103] [DEBUG] [ASRService] ✅ Streaming: 'One, two, three, four, five, six, seven, eight.' (0.11s)
[11:07:39.103] [INFO] [ASRBenchmark] ASR_BENCH t=2848390.533437 session=1 chunk_done index=10 elapsedMs=112 ageMs=7713 samples=112299 rawChars=47 cleanedChars=47 rtf=0.016
[11:07:39.103] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[11:07:39.729] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[11:07:39.729] [INFO] [ASRBenchmark] ASR_BENCH t=2848391.159865 session=1 chunk_start index=11 ageMs=8340 samples=123051 newSamples=10752 audioMs=7691 provider=FluidAudio (Apple Silicon Optimized)
[11:07:39.729] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 123051) using FluidAudio (Apple Silicon Optimized)
[11:07:39.817] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=123051 audioMs=7691 elapsedMs=87 textChars=47 rtf=0.011
[11:07:39.817] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.09s
[11:07:39.817] [DEBUG] [ASRService] ✅ Streaming: 'One, two, three, four, five, six, seven, eight.' (0.09s)
[11:07:39.817] [INFO] [ASRBenchmark] ASR_BENCH t=2848391.247615 session=1 chunk_done index=11 elapsedMs=88 ageMs=8427 samples=123051 rawChars=47 cleanedChars=47 rtf=0.011
[11:07:39.817] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[11:07:40.055] [INFO] [ASRBenchmark] ASR_BENCH t=2848391.485274 session=1 capture_health attempt=2 audioMs=1003 samples=16043 rms=0.000000 peak=0.000000 silent=true inputUID=BuiltInMicrophoneDevice
[11:07:40.454] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[11:07:40.454] [INFO] [ASRBenchmark] ASR_BENCH t=2848391.885068 session=1 chunk_start index=12 ageMs=9065 samples=134657 newSamples=11606 audioMs=8416 provider=FluidAudio (Apple Silicon Optimized)
[11:07:40.454] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 134657) using FluidAudio (Apple Silicon Optimized)
[11:07:40.540] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=134657 audioMs=8416 elapsedMs=86 textChars=47 rtf=0.010
[11:07:40.540] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.09s
[11:07:40.540] [DEBUG] [ASRService] ✅ Streaming: 'One, two, three, four, five, six, seven, eight.' (0.09s)
[11:07:40.540] [INFO] [ASRBenchmark] ASR_BENCH t=2848391.971058 session=1 chunk_done index=12 elapsedMs=86 ageMs=9151 samples=134657 rawChars=47 cleanedChars=47 rtf=0.010
[11:07:40.540] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[11:07:41.057] [INFO] [ASRBenchmark] ASR_BENCH t=2848392.487843 session=1 capture_health attempt=2 audioMs=2005 samples=16043 rms=0.000000 peak=0.000000 silent=true inputUID=BuiltInMicrophoneDevice
[11:07:41.148] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[11:07:41.148] [INFO] [ASRBenchmark] ASR_BENCH t=2848392.578628 session=1 chunk_start index=13 ageMs=9758 samples=145750 newSamples=11093 audioMs=9109 provider=FluidAudio (Apple Silicon Optimized)
[11:07:41.148] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 145750) using FluidAudio (Apple Silicon Optimized)
[11:07:41.233] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=145750 audioMs=9109 elapsedMs=85 textChars=47 rtf=0.009
[11:07:41.233] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.09s
[11:07:41.234] [DEBUG] [ASRService] ✅ Streaming: 'One, two, three, four, five, six, seven, eight.' (0.09s)
[11:07:41.234] [INFO] [ASRBenchmark] ASR_BENCH t=2848392.664215 session=1 chunk_done index=13 elapsedMs=86 ageMs=9844 samples=145750 rawChars=47 cleanedChars=47 rtf=0.009
[11:07:41.234] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[11:07:41.874] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[11:07:41.874] [INFO] [ASRBenchmark] ASR_BENCH t=2848393.304562 session=1 chunk_start index=14 ageMs=10484 samples=157355 newSamples=11605 audioMs=9835 provider=FluidAudio (Apple Silicon Optimized)
[11:07:41.874] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 157355) using FluidAudio (Apple Silicon Optimized)
[11:07:41.959] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=157355 audioMs=9835 elapsedMs=85 textChars=47 rtf=0.009
[11:07:41.959] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.09s
[11:07:41.959] [DEBUG] [ASRService] ✅ Streaming: 'One, two, three, four, five, six, seven, eight.' (0.09s)
[11:07:41.959] [INFO] [ASRBenchmark] ASR_BENCH t=2848393.389794 session=1 chunk_done index=14 elapsedMs=85 ageMs=10569 samples=157355 rawChars=47 cleanedChars=47 rtf=0.009
[11:07:41.959] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[11:07:42.060] [INFO] [ASRBenchmark] ASR_BENCH t=2848393.490514 session=1 capture_health attempt=2 audioMs=3008 samples=16042 rms=0.000000 peak=0.000000 silent=true inputUID=BuiltInMicrophoneDevice
[11:07:42.566] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[11:07:42.566] [INFO] [ASRBenchmark] ASR_BENCH t=2848393.996979 session=1 chunk_start index=15 ageMs=11177 samples=168449 newSamples=11094 audioMs=10528 provider=FluidAudio (Apple Silicon Optimized)
[11:07:42.566] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 168449) using FluidAudio (Apple Silicon Optimized)
[11:07:42.652] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=168449 audioMs=10528 elapsedMs=86 textChars=47 rtf=0.008
[11:07:42.652] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.09s
[11:07:42.653] [DEBUG] [ASRService] ✅ Streaming: 'One, two, three, four, five, six, seven, eight.' (0.09s)
[11:07:42.653] [INFO] [ASRBenchmark] ASR_BENCH t=2848394.083416 session=1 chunk_done index=15 elapsedMs=86 ageMs=11263 samples=168449 rawChars=47 cleanedChars=47 rtf=0.008
[11:07:42.653] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[11:07:43.062] [INFO] [ASRBenchmark] ASR_BENCH t=2848394.493230 session=1 capture_health attempt=2 audioMs=4011 samples=16043 rms=0.000000 peak=0.000000 silent=true inputUID=BuiltInMicrophoneDevice
[11:07:43.275] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[11:07:43.275] [INFO] [ASRBenchmark] ASR_BENCH t=2848394.705884 session=1 chunk_start index=16 ageMs=11885 samples=179713 newSamples=11264 audioMs=11232 provider=FluidAudio (Apple Silicon Optimized)
[11:07:43.275] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 179713) using FluidAudio (Apple Silicon Optimized)
[11:07:43.366] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=179713 audioMs=11232 elapsedMs=91 textChars=47 rtf=0.008
[11:07:43.366] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.09s
[11:07:43.367] [DEBUG] [ASRService] ✅ Streaming: 'One, two, three, four, five, six, seven, eight.' (0.09s)
[11:07:43.367] [INFO] [ASRBenchmark] ASR_BENCH t=2848394.797306 session=1 chunk_done index=16 elapsedMs=91 ageMs=11977 samples=179713 rawChars=47 cleanedChars=47 rtf=0.008
[11:07:43.367] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[11:07:44.007] [DEBUG] [ASRService] 🔄 runStreamingLoop() - calling processStreamingChunk()
[11:07:44.007] [INFO] [ASRBenchmark] ASR_BENCH t=2848395.437435 session=1 chunk_start index=17 ageMs=12617 samples=191489 newSamples=11776 audioMs=11968 provider=FluidAudio (Apple Silicon Optimized)
[11:07:44.007] [DEBUG] [ASRService] Streaming chunk starting transcription (samples: 191489) using FluidAudio (Apple Silicon Optimized)
[11:07:44.065] [INFO] [ASRBenchmark] ASR_BENCH t=2848395.495898 session=1 capture_health attempt=2 audioMs=5013 samples=16043 rms=0.000000 peak=0.000000 silent=true inputUID=BuiltInMicrophoneDevice
[11:07:44.094] [INFO] [ASRBenchmark] ASR_BENCH provider_streaming_done samples=191489 audioMs=11968 elapsedMs=87 textChars=47 rtf=0.007
[11:07:44.094] [DEBUG] [ASRService] Streaming chunk transcription finished in 0.09s
[11:07:44.094] [DEBUG] [ASRService] ✅ Streaming: 'One, two, three, four, five, six, seven, eight.' (0.09s)
[11:07:44.094] [INFO] [ASRBenchmark] ASR_BENCH t=2848395.524948 session=1 chunk_done index=17 elapsedMs=88 ageMs=12705 samples=191489 rawChars=47 cleanedChars=47 rtf=0.007
[11:07:44.094] [DEBUG] [ASRService] 🔄 runStreamingLoop() - processStreamingChunk() returned
[11:07:44.179] [DEBUG] [ContentView] stopAndProcessTranscription called
[11:07:44.179] [INFO] [ContentView] Output route selected: normal
[11:07:44.179] [INFO] [AppBenchmark] APP_BENCH t=2848395.610029 stop_path_enter route=normal
[11:07:44.184] [INFO] [ContentView] Routing decision snapshot | activeMode=dictate | rewrite=false | command=false | overlay=Dictation
[11:07:44.184] [DEBUG] [ContentView] Showing transcription processing state
[11:07:44.184] [INFO] [AppBenchmark] APP_BENCH t=2848395.614818 processing_ui_request status=Transcribing
[11:07:44.184] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848395.614828 manager set_processing_request processing=true overlayVisible=true active=false
[11:07:44.184] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848395.614930 notch text_update status=Transcribing
[11:07:44.184] [INFO] [AppBenchmark] APP_BENCH t=2848395.614952 processing_ui_requested status=Transcribing
[11:07:44.190] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848395.620516 manager processing_show_workitem_execute delayMs=0
[11:07:44.190] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848395.620551 notch set_processing processing=true state=idle bottomVisible=true commandExpanded=false
[11:07:44.190] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848395.620567 bottom_set_processing processing=true
[11:07:44.190] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848395.620575 notch set_processing_forwarded target=bottom
[11:07:44.190] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848395.620578 manager processing_show_workitem_return
[11:07:44.190] [INFO] [AppBenchmark] APP_BENCH t=2848395.620583 asr_stop_call
[11:07:44.190] [INFO] [ASRService] 🛑 STOP() called - beginning shutdown sequence
[11:07:44.190] [INFO] [ASRBenchmark] ASR_BENCH t=2848395.620591 session=1 stop_start ageMs=12800 bufferedSamples=194390
[11:07:44.190] [DEBUG] [ASRService] 📍 Preparing final transcription
[11:07:44.190] [DEBUG] [ASRService] 🚫 Setting isRunning = false...
[INFO] [DirectCoreAudioLifecycle] t=2848395.620640 Direct capture stop begin generation=2 reason=recording_stop retainPrepared=true
[11:07:44.190] [DEBUG] [ASRService] ✅ isRunning disabled
[11:07:44.190] [DEBUG] [ASRService] 👁️ Stopping device monitoring...
[11:07:44.190] [DEBUG] [ASRService] ✅ Device monitoring stopped
[INFO] [DirectCoreAudioLifecycle] t=2848395.636438 Direct capture stop end generation=2 elapsedMs=16 status=0 droppedPackets=0 retained=true
[11:07:44.290] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848395.721193 manager handle_state isRunning=false overlayVisible=true processing=true mode=Dictation
[11:07:44.291] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848395.721220 manager handle_state_return reason=processing_active
[11:07:44.291] [DEBUG] [ASRService] ♻️ Direct audio capture remains prepared
[11:07:44.346] [INFO] [AppBenchmark] APP_BENCH t=2848395.776819 sound_play_request sound=FV_end_0
[11:07:44.346] [INFO] [ASRBenchmark] ASR_BENCH t=2848395.777174 session=1 audio_capture_prepared retained=true phase=prepared generation=2
[11:07:44.346] [DEBUG] [ASRService] ⏳ Awaiting stopStreamingTimerAndAwait()...
[11:07:44.346] [INFO] [ASRBenchmark] ASR_BENCH t=2848395.777192 session=1 streaming_timer_stop begin
[11:07:44.399] [INFO] [AppBenchmark] APP_BENCH t=2848395.829703 sound_play_dispatched sound=FV_end_0 elapsedMs=53
[11:07:44.509] [DEBUG] [ASRService] Streaming transcription task cancelled
[11:07:44.575] [INFO] [ASRBenchmark] ASR_BENCH t=2848396.006078 session=1 streaming_timer_stop end elapsedMs=229 completedChunks=16
[11:07:44.575] [INFO] [ASRBenchmark] ASR_BENCH t=2848396.006112 session=1 stop_streaming_wait elapsedMs=229
[11:07:44.575] [DEBUG] [ASRService] ✅ stopStreamingTimerAndAwait() completed
[11:07:44.575] [INFO] [ASRBenchmark] ASR_BENCH t=2848396.006118 session=1 stop_audio_drained samples=194390 audioMs=12149
[11:07:44.575] [INFO] [ASRBenchmark] ASR_BENCH t=2848396.006144 session=1 silence_gate eligible=false skip=false reason=streaming_preview
[11:07:44.575] [INFO] [ASRBenchmark] ASR_BENCH t=2848396.006153 session=1 stop_ensure_ready skipped=true elapsedMs=0
[11:07:44.576] [DEBUG] [ASRService] Starting transcription with 194390 samples (12.149375 seconds)
[11:07:44.737] [INFO] [ASRBenchmark] ASR_BENCH provider_final_done samples=194390 audioMs=12149 elapsedMs=161 textChars=47 rtf=0.013 fallback=false source=full
[11:07:44.737] [DEBUG] [ASRService] stop(): final transcription finished source=full
[11:07:44.737] [DEBUG] [ASRService] Transcription completed: 'One, two, three, four, five, six, seven, eight.' (confidence: 0.97736925)
[11:07:44.737] [INFO] [ASRService] Final ASR result | provider=FluidAudio (Apple Silicon Optimized) | samples=194390 | textChars=47 | confidence=0.97736925
[11:07:44.737] [INFO] [ASRBenchmark] ASR_BENCH t=2848396.167391 session=1 final_done elapsedMs=161 samples=194390 audioMs=12149 textChars=47 rtf=0.013 streamedChunks=16 source=full
[11:07:44.737] [DEBUG] [ASRService] After post-processing: 'One, two, three, four, five, six, seven, eight.'
[11:07:44.737] [INFO] [ASRBenchmark] ASR_BENCH t=2848396.167568 session=1 stop_end result=success totalMs=547 recordingAgeMs=13347 cleanedChars=47
[11:07:44.737] [INFO] [AppBenchmark] APP_BENCH t=2848396.167639 asr_stop_return elapsedMs=547
[11:07:44.737] [INFO] [ContentView] Stop transcription result | chars=47 | empty=false
[11:07:44.737] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848396.167648 notch text_update status=empty
[11:07:44.745] [DEBUG] [ContentView] Routing transcription through AI post-processing
[11:07:44.745] [INFO] [AppBenchmark] APP_BENCH t=2848396.175408 processing_ui_request status=Refining
[11:07:44.745] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848396.175425 notch text_update status=Refining
[11:07:44.745] [INFO] [AppBenchmark] APP_BENCH t=2848396.175452 processing_ui_requested status=Refining
[11:07:44.823] [DEBUG] [ContentView] processTextWithAI using provider=fluid-1, model=fluid-1
[11:07:44.823] [DEBUG] [ContentView] [PromptTrace][Dictate] Private AI Provider task:
dictationEnhancement
[11:07:44.823] [DEBUG] [ContentView] [PromptTrace][Dictate] Input transcription (Q):
One, two, three, four, five, six, seven, eight.
[11:07:44.823] [DEBUG] [ContentView] [PromptTrace][Dictate] Selected context text:
<none (dictation mode)>
[11:07:44.827] [DEBUG] [PrivateAIProvider] Private provider reused cached runtime model=fluid-1 backend=mlx:/Applications/FluidVoice_169_aug17.app/Contents/Helpers/fluid-intelligence-mlx:/Users/myuser/Library/Application Support/FluidIntelligence/Models/fluid-1-nvfp4-mlx:mtp:/Users/myuser/Library/Application Support/FluidIntelligence/Models/gemma-4-E2B-it-qat-assistant-bf16-mlx-mtp
[11:07:44.827] [INFO] [PrivateAIProvider] Private provider post-processing start backend=MLX model=fluid-1 inputChars=47 maxOutputTokens=256
[11:07:46.475] [INFO] [PrivateAIProvider] Private provider post-processing complete backend=MLX model=fluid-1 totalMs=1624 ttftMs=826 promptTokens=1292 outputTokens=23 totalTokens=1315 tps=28.8
[11:07:46.476] [INFO] [PrivateAIProvider] Private provider post-processing detail backend=MLX mtpEnabled=1 mtpDraftTokens=20 mtpAcceptedTokens=19 mtpAcceptPercent=95 cacheHit=1 reusedTokens=0 prefixTokens=1271 prefillTokens=21 prefillMs=-1 clearMs=-1
[11:07:46.476] [DEBUG] [ContentView] [PromptTrace][Dictate] Model answer (A):
1, 2, 3, 4, 5, 6, 7, 8.
[11:07:46.476] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848397.906332 notch text_update status=empty
[11:07:46.476] [INFO] [ContentView] Dictation AI post-processing finished in 1731ms provider=fluid-1 model=fluid-1 inputChars=47 fallback=false
[11:07:46.476] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848397.906387 notch text_update status=empty
[11:07:46.476] [INFO] [ContentView] Transcription finalized (chars: 24)
[11:07:46.476] [INFO] [AppBenchmark] APP_BENCH t=2848397.906480 transcription_finalized chars=24
[11:07:46.476] [INFO] [AppBenchmark] APP_BENCH t=2848397.906490 text_ready chars=24
[11:07:46.492] [DEBUG] [TranscriptionHistoryStore] Added transcription to history (total: 2770)
[11:07:46.492] [DEBUG] [ContentView] Typing decision → frontmost: FluidVoice, fluidFrontmost: true, editorFocused: false, willTypeExternally: false
[11:07:46.492] [INFO] [AppBenchmark] APP_BENCH t=2848397.922348 overlay_hide_request reason=after_output lifecycle=1
[11:07:46.637] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848398.067796 notch set_processing processing=false state=idle bottomVisible=true commandExpanded=false
[11:07:46.637] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848398.067849 bottom_set_processing processing=false
[11:07:46.637] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848398.067857 notch set_processing_forwarded target=bottom
[11:07:46.637] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848398.067860 manager finish_hide_request
[11:07:46.637] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848398.067879 notch hide_called state=idle bottomVisible=true
[11:07:46.637] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848398.067943 bottom_hide_start windowExists=true
[11:07:46.637] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848398.067960 bottom_hide_animation_start
[11:07:46.691] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848398.121666 bottom_hide_complete elapsedMs=54
[11:07:46.691] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848398.121699 notch hide_return reason=not_visible state=idle notchExists=false
[11:07:46.691] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848398.121704 notch hide_complete target=none elapsedMs=54
[11:07:46.691] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848398.121775 manager finish_hide_complete outcome=hidden elapsedMs=54
[11:07:46.691] [INFO] [AppBenchmark] APP_BENCH t=2848398.121781 overlay_hidden reason=after_output elapsedMs=54
[11:07:56.290] [DEBUG] [ContentView] Mode transition: Optional(FluidVoice.SidebarItem.welcome) → Optional(FluidVoice.SidebarItem.changelog)
[11:07:56.290] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848407.720568 manager set_mode mode=Dictation
[11:07:56.290] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848407.720613 notch set_mode mode=Dictation
[11:08:01.779] [DEBUG] [ContentView] Mode transition: Optional(FluidVoice.SidebarItem.changelog) → Optional(FluidVoice.SidebarItem.preferences)
[11:08:01.779] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848413.210194 manager set_mode mode=Dictation
[11:08:01.779] [INFO] [OverlayBenchmark] OVERLAY_BENCH t=2848413.210241 notch set_mode mode=Dictation