Audio
MOSS-TTS
MOSS-TTS family: open text-to-speech models for voice cloning, dialogue, voice design, and real-time agents
★ 4.2k starsApache-2.0
View guide →Chatterbox — Resemble AI's open-source TTS family: expressive text-to-speech and voice cloning, multilingual V3, low-latency Turbo and tiny Nano models, MIT
Chatterbox is an open-source audio project by resemble-ai (resemble-ai/chatterbox on GitHub). It has 27,079 GitHub stars and 3,645 forks, is released under the MIT license, and was last updated on GitHub on 21 July 2026.
Per the README: pip install chatterbox-tts (developed and tested on Python 3.11); or clone and pip install -e . from source.
Pick the family member for your job — Multilingual V3 for languages and cloning, Turbo for low-latency English agents, Nano for CPU/on-device.
Run the example scripts (example_tts.py, example_vc.py for voice conversion) or the bundled Gradio apps to synthesise speech from text plus a reference voice.
Multilingual V3 (0.5B, many languages), Turbo (350M, one-step decoder for low latency) and Nano (110M, CPU-friendly) cover server to on-device.
Clone a voice from reference audio for TTS, or convert one voice to another with the included VC examples.
Native [laugh], [cough], [chuckle] and similar tags add human texture to generated speech.
The full MIT licence — embed it in commercial products without a separate agreement, unlike custom-licence TTS releases.
The README install.
pip install chatterbox-ttsThe README alternative, in a fresh conda env if you like.
git clone https://github.com/resemble-ai/chatterbox.git
cd chatterbox
pip install -e .Example scripts for TTS and voice conversion ship at the repo root.
python example_tts.py
python example_vc.pyResemble AI (resemble-ai/chatterbox on GitHub), the voice AI company; Chatterbox is their open-source TTS family.
Yes — MIT licensed, including commercial use. Only clone voices you have the right to use.
Multilingual V3 for language coverage and cloning quality, Turbo for low-latency English voice agents, Nano when memory and CPU are the constraint — the README lays out exactly this split.
Against Fish Speech (also on LibreAI), Chatterbox is MIT-licensed and lighter to embed, with a purpose-built low-latency/CPU line; Fish Speech is the larger platform-grade system under a custom community licence. Against XTTS, Chatterbox is newer, actively released and commercially simpler.
pip install chatterbox-tts, then run example_tts.py — or open the Gradio TTS app in the repo.
MOSS-TTS family: open text-to-speech models for voice cloning, dialogue, voice design, and real-time agents
Fish Speech — Fish Audio's expressive open text-to-speech and voice cloning: SOTA-quality TTS you can self-host, under the Fish Audio Community License (check terms)
VoxCPM2: tokenizer-free TTS for multilingual speech, creative voice design and true-to-life cloning
Alibaba's open voice models: expressive, streaming-capable speech with free-form voice design.
IndexTTS: industrial-level, controllable zero-shot text-to-speech with voice cloning
A robust, efficient, low-latency speech-to-text library with advanced voice activity detection, wake word activation and instant transcription.
The most-starred other Audio projects in the LibreAI catalogue — the realistic shortlist if Chatterbox is not the right fit.
利用 AI 大模型和自动化工作流，根据主题或关键词一键生成高清短视频。Generate HD short videos from a topic or keyword with an automated AI workflow.
Robust Speech Recognition via Large-Scale Weak Supervision
1 min voice data can also be used to train a good TTS model! (few shot voice cloning)
Clone a voice from 5 seconds of audio and generate arbitrary speech in real time
Port of OpenAI's Whisper model in C/C++
Free and open — grab the source, star the repo, and start creating.