About Fish Speech
Fish Speech is an open-source text-to-speech project with code, documentation and model weights from the Fish Audio team. It is for developers and researchers who want to run a speech model on their own machines or servers.
What it does
The repository includes the Python codebase, installation guides, a Docker setup and a notebook for running inference. The team describes the project as a state-of-the-art open text-to-speech system, and its README is available in several languages. Running it means downloading the model weights and setting up a GPU environment, either directly or through the Docker files.
Key features
- Python code for speech synthesis
- Docker setup for servers
- Inference notebook included
- Documentation site
- Model weights downloaded separately
Who it's for
Developers, speech researchers and teams that want to self-host a speech model. Users who only want a ready-made web app may prefer a hosted service.
Pricing and good to know
The code and most of the project are released under a Fish Audio community license. Some model weights fall under a separate research license. Read both before using the project commercially. Running the model needs capable hardware.
Highlights
- Generate speech with a Python codebase
- Run the system in Docker containers
- Try inference in a ready notebook
- Download model weights for local use
- Read the docs for install steps
