Overview
ChatTTS is a text-to-speech model built for dialogue, including assistant-style interactions. It supports Chinese and English input, multiple speakers, and controllable prosody such as pauses, laughter, and interjections. The code is available for self-hosted research and development under AGPLv3+. The released model uses a separate CC BY-NC 4.0 license and is marked for academic use, so commercial deployment requires checking the model terms. It can also be integrated into local dialogue prototypes and speech applications for assistants and games alike.