ClovisSTT/transcription
Send an audio file to retrieve its transcription as JSON.
Capabilities
InputWhat you can send
Inputmultipart/form-data : file + model + response_format
OutputWhat the model produces
Outputresponse_format=json
TransportSecure HTTP (HTTPS) — multipart/form-data
Model languagesFrenchEnglishGermanSpanishItalianPortugueseDutchRussianArabicHindiChineseJapaneseKorean
Compatible endpoints
Integration settings
| Setting | Value |
|---|---|
| Endpoint | POST /v1/audio/transcriptions |
| Model | ClovisSTT/transcription |
| Response format | json — set explicitly |
| Transcript | transcript.text.strip() with the Python SDK |
| Audio format tested | MP3 |
| Language codes tested | fr, en — optional language field |
The gateway rejects requests that omit response_format with HTTP 400. With response_format="text", it returns nested JSON. Use json to access the transcript directly.
Transcribe an existing recording
Use ClovisSTT/transcription for completed recordings such as meetings, interviews, podcasts or voice messages. The JSON response contains the transcript in its text field. Speaker labels, structured documents and advanced transcript cleaning are not part of the documented contract.
The route is available on staging. The HTTP guide describes the parameters and formats verified on this environment.
See the HTTP integration guide for Python and cURL examples. To send audio continuously, use ClovisSTT/realtime.