RepoPicks Radar

Updated every day around 23:30 UTC

Guides

ChatTTS

2noise/ChatTTSReviewed by AI

Checked by RepoPicks Radar on

What is ChatTTS?

A text-to-speech model made for dialogue: you type text and it reads it aloud with pauses, laughter and several different voices, in English or Chinese. It comes with a simple web page you start yourself and a command line example, and it is also usable as a Python library. The project calls itself the algorithm plus simple examples and sends end users to community projects built on it. You need Python, a graphics card with at least 4 GB of memory for a 30-second clip, and the published model is for study and research, not commercial work.

ChatTTS is free and open source (AGPL-3.0) and can replace ElevenLabs and Murf.

How easy is it to install ChatTTS?

C · install 1 or 2 tools · Web

  • Windows D
  • macOS D
  • Linux D
  • Web C

The easiest way in is the web: the repository ships a Google Colab notebook you open in the browser and run cell by cell, so no install, which is a C. On Windows, Mac and Linux it is a D: you clone the repository, build a Python environment, install the requirements and run `python examples/web/webui.py`, and the FAQ asks for a graphics card with at least 4 GB of memory; some extra packages are Linux-only and Macs have no documented graphics support.

ChatTTS on GitHub

Stars
39,998on Nov 10, 2026
Downloads
—all releases
Growth
+23stars in 7 days

Stars over time

39,98539,99039,99540,000 Nov 5Nov 6Nov 7Nov 8Nov 9Nov 10 39,98539,99039,99540,000 Nov 5Nov 8

Facts

Runs on
Windows, macOS, Linux, Web
You need to install
Python, Graphics card
To build from source
Python, Graphics card
Disk space
950 MB–4.3 GB · estimate
Accounts or keys
None
Cost
Free · No price. The code is AGPLv3+, but the model files are CC BY-NC 4.0: the README says they are "for academic purposes only" and "should not be used for any commercial or illegal purposes".
License
AGPL-3.0 (Copyleft)
Can replace
ElevenLabs, Murf
Last change on GitHub
Apr 10, 2026

ChatTTS risk signals (5)

  • Attentionmodel license forbids commercial use

    README.md: "The released model is for academic purposes only" and "The model is published under `CC BY-NC 4.0` license. It is intended for educational and research use, and should not be used for any commercial or illegal purposes." The code itself is AGPLv3+, which is a different licence from the model you download.

    AI review

  • Inforeleased model deliberately degraded

    README.md disclaimer: "we added a small amount of high-frequency noise during the training of the 40,000-hour model, and compressed the audio quality as much as possible using MP3 format, to prevent malicious actors from potentially using it for criminal purposes." The README also says the public model is a 40,000-hour version without fine-tuning, while the main model was trained on 100,000+ hours.

    AI review

  • Infooutput quality varies

    README FAQ: "Model stability is not good enough, with issues such as multi speakers or poor audio quality... One can try multiple samples to find a suitable result." Only `[laugh]`, `[uv_break]` and `[lbreak]` can be controlled directly, English is marked experimental in the example, and only English and Chinese are supported.

    AI review

  • Infovoice imitation possible

    The roadmap marks "Open-source DVAE encoder and zero shot inferring code" as done, and the code samples a speaker vector you can save and reuse, so the project can produce voices that imitate a sample. The authors write that they "internally trained a detection model and plan to open-source it in the future".

    AI review

  • Infodevelopment slowed

    The last commit and the last release (v0.2.5) in the checked snapshot are both from 2026-04-10, 208 days before the check on 2026-11-04, and roadmap items such as multi-emotion control and ChatTTS.cpp are still open with 63 open issues.

    AI review

Evidence is shown in its original language.

What Radar could not check about ChatTTS

  • How the voices really sound to you, especially in English
  • Whether it runs on a Mac or on a computer without a dedicated graphics card, and how slowly
  • Real disk space taken by the downloaded model files
  • Whether the Google Colab notebook still works with current versions of the libraries
  • How the non-commercial model licence applies to what you want to do with the audio

Similar projects

The RepoPicks app

Coming soon: install in 1 click with the RepoPicks app.

Before you download

See the release page on GitHub