fish-speech
fishaudio/fish-speech · Python · Other
Multilingual open TTS with an active community
Fish Speech is a full TTS project with Docker options and strong Chinese/English coverage. Check the exact license of the weights you download before shipping.
Score
Setup at a glance
git clone https://github.com/fishaudio/fish-speech && pip install -e . Deep dive
Fish Speech is a multilingual TTS and voice-cloning toolkit with one of the strongest open-source communities in speech: a short reference clip is enough to clone a voice, and the same stack serves dozens of languages.
It exposes three inference paths — command line, WebUI and an OpenAI-compatible server — plus a Docker setup, so the same model fits a quick experiment, a demo and a production service.
Under the hood it combines VQGAN, VALLE and VITS-style components with a Llama backbone, and the project publishes installation, inference and server documentation for each path.