transcription-debugger

Diagnose transcription failures by verifying the AudioFileReady event fires during recording, confirming the Whisper executable runs without hanging, and inspecting VoiceLite logs for error patterns. Trace the event chain from recording through transcription to isolate where the failure occurs.

Silent failures occur when the audio processing pipeline completes without errors but produces no text output. Enable comprehensive logging across all builds and validate event subscriptions to detect when critical lifecycle events fail to trigger or propagate.

Manually execute the whisper.exe subprocess with test audio files and inspect its output and error streams. This isolates whether transcription failures stem from the Whisper model, file integrity issues, or event-chain problems upstream.

Verify the Whisper executable exists and is accessible, confirm audio files are valid WAV or MP3 format with proper permissions, check system resources, and review Whisper process logs for model loading or decoding errors.

Yes. The debugging scope covers recording and transcription across the full VoiceLite pipeline regardless of input source. Diagnostic guidance applies to any scenario where voice typing produces no output or fails silently.

This typically indicates the AudioFileReady event isn't firing or isn't triggering the Whisper subprocess. Validate event subscriptions, check the event chain for broken links, and verify the transcription handler is properly registered in your lifecycle configuration.