Plain-AyI
← All projects
Speech & audio

CosyVoice

Reads text aloud in natural-sounding speech in many languages, and can copy a voice from a short sample.

Research-backed

What it does

CosyVoice is a family of text-to-speech models: give it text and it speaks it aloud. The latest version, Fun-CosyVoice 3.0, covers nine common languages, including Chinese, English, Japanese, Korean, German, Spanish, French, Italian and Russian, plus many Chinese dialects and accents.

It can do "zero-shot voice cloning": imitate a voice from a short recording without extra training, even across languages. You can steer the language, dialect, emotion, speed and volume with instructions, correct how particular words are pronounced, and stream audio as it's generated. Its makers publish test results comparing it with other speech models.

The project includes the models, training code and tools for running it as a service.

Who it's for

Developers and researchers building voice apps, audiobooks, assistants or dubbing tools.

Can I use it without coding?

Developers only. Needs Python, a Conda environment and downloading the models; a local web demo is available once it's set up. New to this? Start here.

Can I run it?

Cost
FreeNothing to pay or sign up for.
Runs on
Linux
Computer
Works best on a powerful computer or one with a graphics card (GPU)
Privacy
Runs on your own device: your data stays with you
API key
Not needed

Try it

Using Claude Code, Cursor or another AI coding assistant? Paste the prompt and it will walk you through the setup.New to this?

Before you install

  • Only download from the official GitHub page. Copies with similar names exist for many popular projects.
  • License (Apache-2.0): Permissive: free to use, change and share, including for business.
  • More on staying safe

Where people are talking about it