The local tts app has an API server, and it would be nice to have the ability to integrate with that server.
In particular, the advantage is that you can set the voice you want to communicate through example voices, and that very fast inferences are possible.
Originally created by @codeMonkey-shin on GitHub (Nov 21, 2024).
Original GitHub issue: https://github.com/open-webui/open-webui/issues/7135
Originally assigned to: @tjbck on GitHub.
https://github.com/fishaudio/fish-speech
The local tts app has an API server, and it would be nice to have the ability to integrate with that server.
In particular, the advantage is that you can set the voice you want to communicate through example voices, and that very fast inferences are possible.
Wondering what is the latest status here as I see #11230 has been closed. Any further work or discussion needed? Thanks.
<!-- gh-comment-id:2906903339 -->
@codgician commented on GitHub (May 24, 2025):
Wondering what is the latest status here as I see #11230 has been closed. Any further work or discussion needed? Thanks.
For those looking for a lightweight, self-hosted multilingual TTS option that works with Open WebUI's existing OpenAI-compatible TTS integration:
piper-plus provides an OpenAI-compatible /v1/audio/speech endpoint out of the box (via Docker or FastAPI), supporting 6 languages (JA/EN/ZH/ES/FR/PT) in a single ~75MB ONNX model — no GPU required.
Quick setup with Docker:
docker compose -f docker/python-inference/docker-compose.yml up
Then point Open WebUI's TTS settings to http://<host>:8000/v1/audio/speech.
Key features:
Runs on CPU (Raspberry Pi-capable)
6-language single model, no eSpeak-ng (MIT license)
<!-- gh-comment-id:4230568904 -->
@ayutaz commented on GitHub (Apr 12, 2026):
For those looking for a lightweight, self-hosted multilingual TTS option that works with Open WebUI's existing OpenAI-compatible TTS integration:
[piper-plus](https://github.com/ayutaz/piper-plus) provides an OpenAI-compatible `/v1/audio/speech` endpoint out of the box (via Docker or FastAPI), supporting 6 languages (JA/EN/ZH/ES/FR/PT) in a single ~75MB ONNX model — no GPU required.
Quick setup with Docker:
```
docker compose -f docker/python-inference/docker-compose.yml up
```
Then point Open WebUI's TTS settings to `http://<host>:8000/v1/audio/speech`.
Key features:
- Runs on CPU (Raspberry Pi-capable)
- 6-language single model, no eSpeak-ng (MIT license)
- Voice cloning from reference audio
- ~75MB FP16 ONNX model
Demo: https://huggingface.co/spaces/ayousanz/piper-plus-demo
Models: https://huggingface.co/ayousanz/piper-plus-base
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Originally created by @codeMonkey-shin on GitHub (Nov 21, 2024).
Original GitHub issue: https://github.com/open-webui/open-webui/issues/7135
Originally assigned to: @tjbck on GitHub.
https://github.com/fishaudio/fish-speech
The local tts app has an API server, and it would be nice to have the ability to integrate with that server.
In particular, the advantage is that you can set the voice you want to communicate through example voices, and that very fast inferences are possible.
@tjbck commented on GitHub (Nov 21, 2024):
Looks great, will investigate shortly!
@a1558878 commented on GitHub (Jan 17, 2025):
What's the result?
@codgician commented on GitHub (May 24, 2025):
Wondering what is the latest status here as I see #11230 has been closed. Any further work or discussion needed? Thanks.
@ayutaz commented on GitHub (Apr 12, 2026):
For those looking for a lightweight, self-hosted multilingual TTS option that works with Open WebUI's existing OpenAI-compatible TTS integration:
piper-plus provides an OpenAI-compatible
/v1/audio/speechendpoint out of the box (via Docker or FastAPI), supporting 6 languages (JA/EN/ZH/ES/FR/PT) in a single ~75MB ONNX model — no GPU required.Quick setup with Docker:
Then point Open WebUI's TTS settings to
http://<host>:8000/v1/audio/speech.Key features:
Demo: https://huggingface.co/spaces/ayousanz/piper-plus-demo
Models: https://huggingface.co/ayousanz/piper-plus-base