Body_audio_transcriptions_v1_audio_transcriptions_post

fileFile (string)required

Audio file to transcribe.

modelModel (string)required

Mounted speech-to-text model id to serve.

language object

Optional input language hint.

anyOf

• string
• null

string

prompt object

Optional transcription prompt or context.

anyOf

• string
• null

string

response_formatResponse Format (string)

Transcription response format.

Possible values: [json, text, verbose_json, srt, vtt, ndjson]

Default value: json

temperature object

Optional model-specific sampling temperature.

anyOf

• number
• null

number

streamStream (boolean)

Return typed SSE transcript events as model deltas arrive. With response_format=ndjson, retain progressive NDJSON framing.

Default value: false

max_tokens object

Optional model-specific maximum token budget.

anyOf

• integer
• null

integer

chunk_duration object

Optional model-specific audio chunk duration.

anyOf

• number
• null

number

frame_threshold object

Optional model-specific frame threshold.

anyOf

• integer
• null

integer

context object

Optional model-specific context text.

anyOf

• string
• null

string

prefill_step_size object

Optional model-specific prefill step size.

anyOf

• integer
• null

integer

text object

Optional model-specific text prefix.

anyOf

• string
• null

string

word_timestampsWord Timestamps (boolean)

Whether to request word timestamp metadata when supported.

Default value: false

timestamp_granularities object

Comma-separated or JSON list of timestamp granularities.

anyOf

• string
• null

string

Body_audio_transcriptions_v1_audio_transcriptions_post

{
  "file": "string",
  "model": "string",
  "language": "string",
  "prompt": "string",
  "response_format": "json",
  "temperature": 0,
  "stream": false,
  "max_tokens": 0,
  "chunk_duration": 0,
  "frame_threshold": 0,
  "context": "string",
  "prefill_step_size": 0,
  "text": "string",
  "word_timestamps": false,
  "timestamp_granularities": "string"
}