Skip to content

ChatTTS.com

A speech model built for natural dialogue

Visit official site

What ChatTTS is

The model is described as a voice generation model designed for dialogue scenarios, covering conversation tasks of large language model assistants and conversational audio and video introductions. It supports Chinese and English after training on roughly 100,000 hours of data in both languages.

What you can do with it

  • Generate a speech file from text alone
  • Use Chinese and English within one model
  • Handle dialogue tasks of large language model assistants
  • Produce speech for conversational video introductions

Getting started

  • Download the code from GitHub
  • Install the required dependencies such as torch and ChatTTS
  • Import the needed libraries including torch, ChatTTS and Audio from IPython.display
  • Create a ChatTTS instance and load the pretrained model

Model notes

  • The feature list includes a plan to open-source the trained base model
  • Work on model controllability is described as ongoing
  • Watermarking and integration into large language models are listed as aims

Pros & cons

✓ What we like

  • Text alone is enough to generate a speech file
  • Dialogue tasks fit the intended design of the model

! What to watch out for

  • A code download and dependency install are part of the documented setup
  • Watermarking and controllability are described as ongoing work

FAQ

Which languages are named?

The material names Chinese and English, trained on roughly 100,000 hours of data in both languages.

What is planned for the model?

The feature list mentions a plan to open-source the base model, add watermarks and integrate it into large language models.

Last reviewed: 2026-11-15

More AI audio tools tools

View all →

How we review