OpenAI-Compatible TTS

Use this endpoint when a client expects the OpenAI POST /v1/audio/speech contract. It is intended for integrations such as Tavo's Custom OpenAI Protocol TTS provider.

For a native Fish Audio integration with the full request contract, use Sync HTTP.

Endpoint

POST /v1/audio/speech
Authorization: Bearer YOUR_API_KEY
Content-Type: application/json

Full URL:

https://fishaudio.org/v1/audio/speech

Keep the API key on a trusted server or in the third-party provider's protected API credential field. Create a separate key for each customer or integration so it can be monitored and revoked independently.

Request

{
  "model": "fishaudio-s21pro-flash",
  "input": "Hello from Fish Audio.",
  "voice": "00a1b221-6137-4b73-ad62-b0cbce134167",
  "response_format": "mp3",
  "speed": 1
}

Get voice ids from GET /api/open/v1/voices. For a quick connection test, use the public system voice 00a1b221-6137-4b73-ad62-b0cbce134167. Use a supported Fish Audio TTS model id; OpenAI model names such as tts-1 are not aliases for Fish Audio models.

Some third-party clients discover voices through an OpenAI-compatible voice list. They can call:

GET /v1/audio/voices
Authorization: Bearer YOUR_API_KEY

Use a returned voices[].id value as the voice in a speech request.

Response

A successful request returns binary audio:

HTTP/1.1 200 OK
Content-Type: audio/mpeg
X-OpenAPI-Quota-Remaining: 987988
X-OpenAPI-Credits-Used: 12

<binary audio data>

The client should save or play the response according to Content-Type. Errors are JSON responses and use the same authentication, quota, and generation behavior as the native sync endpoint.

curl Example

curl -X POST "https://fishaudio.org/v1/audio/speech" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "fishaudio-s21pro-flash",
    "input": "Hello from Fish Audio.",
    "voice": "00a1b221-6137-4b73-ad62-b0cbce134167",
    "response_format": "mp3",
    "speed": 1
  }' \
  --output speech.mp3

Connect Tavo

For API key, Base URL, model, voice, and character binding setup, follow the dedicated Fish Audio TTS for Tavo guide.

Errors