POST /v1/audio/transcriptions
Speech-to-text endpoint.
Inference APIs
Transcribe audio via multipart form upload and return the upstream JSON body unchanged.
POST/v1/audio/transcriptions
Request
- Auth: tenant auth when enabled; otherwise dev mode is open
- Content-Type:
multipart/form-data - Max body size: 64 MiB
Request fields
| Field | Type | Required | Notes |
|---|---|---|---|
model | form field | yes | Gateway requires it to resolve the route |
file | form field | yes for upstream transcription APIs | Forwarded unchanged |
language | form field | no | Forwarded unchanged |
prompt | form field | no | Forwarded unchanged |
response_format | form field | no | Forwarded unchanged |
temperature | form field | no | Forwarded unchanged |
| other upstream fields | form field | no | Forwarded unchanged |
Response schema
| Field | Type | Required | Notes |
|---|---|---|---|
text | string | yes | Upstream transcription text |
Errors
| HTTP | Code | When |
|---|---|---|
400 | invalid_request_error | Invalid content type, missing boundary, malformed multipart body, or missing model |
400 | provider_not_configured | Provider missing from config or unavailable |
400 | invalid_request_error | Provider does not support audio transcriptions |
502 | upstream_error | Upstream request failed or response could not be read |
Notes
- The gateway forwards the multipart body as-is, including the boundary.
- Response content is raw upstream JSON.
- The response example is copyable, but the production payload is just upstream JSON text.