POST /v1/audio/speech

Text-to-speech endpoint.

Inference APIs

Synthesize audio from text and stream back the provider's binary audio response with the right content type.

POST/v1/audio/speech

Request

Auth: tenant auth when enabled; otherwise dev mode is open
Content-Type: application/json
Max body size: 256 KiB

Request fields

Field	Type	Required	Notes
`model`	string	yes	Alias or direct provider string
`input`	string	yes	Text to synthesize
`voice`	string	yes	Voice name
`response_format`	string	no	Maps to audio content type; unsupported or omitted values default to `audio/mpeg`
`speed`	number	no	Passed upstream when present

Response handling

Field	Type	Required	Notes
response body	binary	yes	Upstream audio payload
`Content-Type`	header	yes	Preserved from upstream when set, otherwise derived from `response_format`

Errors

HTTP	Code	When
`400`	`invalid_request_error`	Malformed JSON or missing `model`, `input`, or `voice`
`400`	`provider_not_configured`	Provider missing from config or unavailable
`400`	`invalid_request_error`	Provider does not support speech
`502`	`upstream_error`	Upstream request failed

Notes

The response is binary, not JSON.
The gateway preserves the upstream content type when possible.
Format mapping falls back to audio/mpeg when no explicit match exists.

POST /v1/images/generations

Image generation endpoint.

POST /v1/audio/transcriptions

Speech-to-text endpoint.