OpenWebUI Integration Guide

Last updated: 2025-02-02

Setup -> Setting Base URL

• If you are running OpenWebUI as a Docker container, you need to change the base URL in OpenWebUI settings to
  • http://host.docker.internal:8880/v1
• If not, you will need to point it at
  • http://localhost:8880/v1

This ensures that OpenWebUI can communicate with the Kokoro FastAPI wrapper running in another Docker container.

---

Configure TTS

Using Kokoro as the Model

You can use Kokoro as the model in OpenWebUI. Here's how you can integrate it:

Set the Model

• In OpenWebUI, go to the settings or configuration section.
• Set the model to kokoro, api key to 'not-needed'

Use Mapped OpenAI Voices

• OpenWebUI supports mapped OpenAI voices. You can use any of the mapped voices directly.
  • e.g. alloy,echo,fable,onyx,nova,shimmer
• You can see how it's currently mapped in api/src/core/openai_mappings.json, and modify (if running from compose or uv), or see the current mapping (at time of this post) as:

    "voices": {
        "alloy": "am_adam",
        "ash": "af_nicole",
        "coral": "bf_emma",
        "echo": "af_bella",
        "fable": "af_sarah",
        "onyx": "bm_george",
        "nova": "bf_isabella",
        "sage": "am_michael",
        "shimmer": "af_sky"
}

Use Direct Voice Names

• If you prefer, you can use the direct voice names provided by Kokoro.
• Set the voice parameter to the desired voice name (without the file extension)
• e.g `af_nicole','am_michael', etc