ChatTTS

ChatTTS is a cutting-edge conversational text-to-speech (TTS) model purpose-built for dialogue scenarios such as chatbots and virtual assistants. Instead of the flat, robotic output of traditional TTS engines, it transforms plain text into dynamic, natural-sounding speech that feels like a real conversation, making interactions with your applications far more engaging and human.

The model supports both English and Chinese, so you can serve users across languages with a single solution. It is trained on extensive data—more than 100,000 hours for the full version and 40,000 hours for the open-source version—which gives it the depth needed to deliver expressive, lifelike speech.

A standout strength of ChatTTS is its fine-grained control over prosodic features. You can shape details like laughter, pauses, and interjections, adding natural rhythm and emotion to the voice. This makes spoken responses feel warmer and more authentic, whether you are building a friendly assistant, an interactive agent, or any voice-driven dialogue experience.