Voice

Voice

Whisper-based transcription and streaming voice sessions, available behind the voice feature flag.

Transcription

Voice transcription is powered by whisper-rs. The functionality is feature-gated and compiled with --features voice.

Endpoints

The voice API includes:

• /voice/transcribe
• /voice/stream/{id}
• /voice/stream/{id}/chunk

Streaming sessions

Streaming mode uses voice sessions that receive audio incrementally and expose transcription results over the session lifecycle.