Skip to main content
For coding agents, discover the current recommended STT shortlist first with GET /v1/models?recommended_for=stt, then send the selected model explicitly to this endpoint.

Request Body

Synchronous request timeout: This non-chat endpoint waits for the routed model to finish. Large inputs, long audio, or large batches can exceed common 30s client defaults, so set your HTTP client timeout to at least 120s.
file
required
Audio file to transcribe. Supported formats: flac, mp3, mp4, mpeg, mpga, m4a, ogg, wav, webm.
string
default:"whisper-1"
Model to use. Currently only whisper-1 is supported.
string
Language of the audio in ISO-639-1 format (e.g., en, zh, ja).
string
Optional text to guide the model’s style or continue a previous segment.
string
default:"json"
Output format: json, text, srt, verbose_json, vtt.
number
default:"0"
Sampling temperature (0 to 1).
array
Timestamp granularity: word and/or segment. Requires verbose_json.

Response

string
The transcribed text.
For verbose_json:
string
Always transcribe.
string
Detected language.
number
Audio duration in seconds.
array
Transcription segments with timestamps.
array
Word-level timestamps (if requested).

Translation

To translate audio to English, use the translations endpoint: