Voice input with voice response (full voice-to-voice)

POST 
/chat/voice-response

Full voice-to-voice pipeline: STT (Whisper) → AI Agent → TTS (OpenAI) → Audio Response.

Streaming Mode: Add stream: "true" in form data to receive SSE stream including:

• transcription event (STT result)
• agent_thinking, token, agent_response events
• tts_start, tts_complete events with base64 audio
• stream_end with full performance metrics

Note: This endpoint uses a different base path: /api/chat/voice-response

Request

Responses

200

Standard Mode: JSON response with transcription, AI response, and audio.

Streaming Mode: SSE stream with all events including TTS audio.