Module transcribe 

omni-dev voice transcribe — feed a 16 kHz mono WAV file through the configured crate::voice::Transcriber and emit JSONL events to stdout (markdown when stdout is a tty).

WAV validation is delegated to crate::voice::VecAudioInput::from_wav_path — non-16 kHz, non-mono, non-16-bit-PCM files error with a descriptive message pointing at voice capture as the source of normalised audio.

Structs

TranscribeCommand

Transcribes a 16 kHz mono WAV file to JSONL or markdown.

Enums

OutputFormatArg

clap value enum matching OutputFormat.

Constants

DEFAULT_SPEAKER_THRESHOLD

Default cosine-similarity threshold for --speaker filtering.