Connect Fish Audio TTS to Open WebUI

Open WebUI can use an OpenAI-compatible endpoint for text-to-speech. After setup, click the speaker icon beside an AI response to hear it with a Fish Audio voice.

Prepare the API settings

Create a dedicated key in API management, then use:

Configure the admin interface

Sign in as an administrator and open:

Admin Panel → Settings → Audio → Text-to-Speech

Select the OpenAI engine and enter the Base URL, API key, model, and voice. Do not append /audio/speech to the Base URL.

Configure environment variables

For a Docker or server deployment, set:

AUDIO_TTS_ENGINE=openai
AUDIO_TTS_OPENAI_API_BASE_URL=https://fishaudio.org/v1
AUDIO_TTS_OPENAI_API_KEY=YOUR_API_KEY
AUDIO_TTS_MODEL=fishaudio-s21pro-flash
AUDIO_TTS_VOICE=00a1b221-6137-4b73-ad62-b0cbce134167

Restart Open WebUI after updating the environment. Settings previously saved in the admin interface may override startup values, so verify the displayed configuration.

Test speech

Start a chat, generate a short response, and click its speaker icon. After it plays successfully, replace the test voice with your preferred Fish Audio Voice ID.

Troubleshooting

• 401: the API key is missing, contains whitespace, or was revoked.
• 402: API credits are insufficient; membership credits are a separate balance.
• Invalid model or voice: use a Fish Audio model and Voice ID, not tts-1 or alloy.
• No speaker icon: confirm that the TTS engine is enabled and reload the chat.
• Long replies stop early: enable punctuation or paragraph response splitting in Audio settings.

Related documentation

• OpenAI-Compatible TTS
• API pricing
• Official Open WebUI TTS guide