Realtime Speech-to-Text

GET 
wss://api.knoxstudio.ai/v1/speech-to-text/realtime

Low-latency streaming transcription (~150ms). Default model scribe_v2_realtime at $0.44/hour. GET and POST accept the WebSocket upgrade. Plain HTTP returns 400. Do not put credentials in JSON bodies; Knox strips xi-api-key.

Request

Query Parameters

Name	Type	Required	Description
model_id	String	No	Default scribe_v2_realtime. Must be a realtime STT model.
language_code	String	No	Hint for recognition.
audio_format	String	No	pcm_8000, pcm_16000, pcm_22050, pcm_24000, pcm_44100, pcm_48000, ulaw_8000.
commit_strategy	String	No	manual or vad.
enable_logging	Boolean	No	Default true.
include_timestamps	Boolean	No	Include word timings on transcripts.
include_language_detection	Boolean	No	Include language detection events.

Auth: Authorization: Bearer, xi-api-key, or (browser) ?authorization= / ?api_key=. Upgrade: websocket is required.

Request Headers

Name	Type	Required	Description
Authorization	String	Yes	Bearer <token> Knox API key. xi-api-key is also accepted.
Upgrade	String	Yes	Must be websocket.

Client messages

{
  "message_type": "input_audio_chunk",
  "audio_base_64": "<base64 pcm>",
  "commit": false,
  "sample_rate": 16000
}

Set commit: true (or send a commit message) when commit_strategy=manual. VAD commits automatically.

Server messages

• partial_transcript — interim text for the current segment
• committed_transcript — stable text for that segment

Billing starts after committed audio is observed.

Example

const ws = new WebSocket(
  "wss://api.knoxstudio.ai/v1/speech-to-text/realtime?model_id=scribe_v2_realtime&audio_format=pcm_16000&commit_strategy=vad"
);
ws.onopen = () => {
  ws.send(JSON.stringify({
    message_type: "input_audio_chunk",
    audio_base_64: pcmChunk,
    commit: false,
    sample_rate: 16000
  }));
};

Error Responses

Validation uses HTTP 422:

{
  "detail": [
    {
      "loc": ["body", "field"],
      "msg": "…",
      "type": "value_error"
    }
  ]
}

Other errors use { "detail": "message" }.

Status	When
400	WebSocket path called without Upgrade: websocket.
401	Missing or invalid API key.
402	Insufficient credits.
403	Token is not allowed to use this model.
404	Resource not found.
422	Validation error.
429	Rate limit exceeded.
500	Internal server error.