ChatTTS
Generative speech model for daily dialogue
ChatTTS is a generative speech model optimized for daily dialogue scenarios, supporting both English and Chinese languages. Trained with over 100,000 hours of data, it features fine-grained control over prosodic elements such as laughter, pauses, and interjections, enhancing natural and expressive speech synthesis. The project is open-source and available on GitHub, with a 40,000-hour pretrained model accessible on HuggingFace. It includes functionalities for conversational TTS, multi-speaker support, and manual prosody adjustments. The repository is intended for academic and research purposes only.