ChatTTS

// Overview

ChatTTS is a generative text-to-speech model designed for dialogue scenarios such as LLM assistants. Instead of reading text in a flat, robotic way, it aims to produce speech that sounds natural and expressive in a back-and-forth conversation.

The model supports English and Chinese, and can render multiple speakers so interactive conversations feel more lifelike. The open-source release on Hugging Face is a 40,000-hour pre-trained base model, intended for academic and research use under a non-commercial license.

// What it does

• Conversational TTS optimized for dialogue, with support for multiple speakers in interactive conversations
• Fine-grained control over prosody, including laughter, pauses, and interjections via in-text tokens
• Better prosody than most open-source TTS models, with pretrained weights provided for further research
• Sentence-level and word-level control using tokens like [oral_2], [laugh], [uv_break], and [break_6]
• Streaming audio generation and zero-shot inference with the open-sourced DVAE encoder
• Sample random speaker embeddings from a Gaussian and save them to recover the same timbre later

// Getting started

ChatTTS is distributed as a Python package on PyPI. Install it, load the model, and call infer() with your text to produce audio. The example below mirrors the basic usage in the project README.

pip install ChatTTS

import ChatTTS
import torch
import torchaudio

chat = ChatTTS.Chat()
chat.load(compile=False) # Set to True for better performance

texts = ["PUT YOUR 1st TEXT HERE", "PUT YOUR 2nd TEXT HERE"]

wavs = chat.infer(texts)

for i in range(len(wavs)):
    torchaudio.save(f"basic_output{i}.wav", torch.from_numpy(wavs[i]).unsqueeze(0), 24000)

python examples/cmd/run.py "Your text 1." "Your text 2."
python examples/web/webui.py

Commands and code are distilled from the project's own documentation — always check the official repo for the latest.

// When to use it

• Giving an LLM assistant or chatbot a natural, expressive spoken voice for dialogue
• Research and academic experiments on prosody and conversational speech synthesis
• Generating multi-speaker English and Chinese audio with controllable laughter, pauses, and intonation

// How ChatTTS compares

ChatTTS alongside other open-source audio, music & voice tools AI/TLDR tracks, ranked by GitHub stars.