Devtoolwatch BoardChatterbox review
Chatterbox is a family of open-source text-to-speech and voice cloning models by Resemble AI, with English and multilingual variants.
TL;DR
Chatterbox is a family of open-source text-to-speech models from Resemble AI, installed as a Python package. It suits developers building voice agents, narration, localization, or on-device speech. It offers zero-shot voice cloning, paralinguistic tags like [laugh] in Turbo and Nano, emotion exaggeration control, and built-in Perth watermarking on every generated file.
- 27.5kGitHub stars
- +285new stars, week of Oct 25
- 73.2kdownloads a week, +44% vs 4-week avg
- #3in Voice and realtime this week
Chatterbox at a glance
- Category
- Voice and realtime
- Maintained by
- Resemble AI (organization, @resemble-ai on GitHub)
- License
- MIT
- Language
- Python
- Repository
- github.com/resemble-ai/chatterbox
- Website
- resemble-ai.github.io/chatterbox_demopage
- PyPI package
- chatterbox-tts 0.1.7
- Latest release
- chatterbox-tts 0.1.7 on PyPI, Mar 26, 2026
- First published
- Apr 23, 2025
What Chatterbox does
- Chatterbox generates speech from text and clones a voice from a reference audio clip without finetuning.
- Chatterbox Multilingual V3 supports 23 languages, including Arabic, Hindi, Japanese, Chinese, and Swahili, with cross-language voice cloning.
- Chatterbox-Turbo is a 350M-parameter English model for low-latency voice agents, using a one-step speech-token-to-mel decoder.
- Chatterbox-Nano is a 110M-parameter English model for on-device and CPU inference, described as 3x faster than realtime on 8 CPU cores.
- Chatterbox Turbo and Nano support paralinguistic tags such as [cough], [laugh], and [chuckle] inside the input text.
- Chatterbox exposes exaggeration and cfg_weight settings to control expressiveness and pacing of generated speech.
- Chatterbox embeds an imperceptible Perth watermark in every generated audio file, which a perth script can detect.
How it works
Chatterbox runs as a Python library. You load a model class such as ChatterboxTurboTTS, ChatterboxTTS, or ChatterboxMultilingualTTS with from_pretrained on a device like cuda, cpu, or mps. You then call generate with text and an optional reference clip, and save the waveform with torchaudio. Model weights are hosted on Hugging Face.
How to install and start using Chatterbox
With pip
pip install chatterbox-ttsFrom source
git clone https://github.com/resemble-ai/chatterbox.gitRequirements
- Python 3.10 or newer; developed and tested on Python 3.11 on Debian 11.
- A CUDA GPU, CPU, or Apple MPS device for inference.
- A reference audio clip for voice cloning with Turbo and Nano.
First steps
- Install the package with pip install chatterbox-tts in a Python 3.11 environment.
- Load a model, for example ChatterboxTurboTTS.from_pretrained(device="cuda"), or pass nano=True for Chatterbox-Nano.
- Call model.generate with your text and audio_prompt_path set to a reference clip of about 10 seconds.
- Save the output with ta.save("test-turbo.wav", wav, model.sr).
Who Chatterbox is for
- Developers building low-latency English voice agents who want a self-hosted model.
- Teams doing localization or cross-language voice cloning across many languages.
- On-device or CPU deployments with tight memory and latency budgets, using Chatterbox-Nano.
- Creators of videos, games, and narration who want emotion exaggeration control.
Limitations to know
- Chatterbox-Turbo and Chatterbox-Nano support English only; other languages need the Multilingual or Single Language Pack models.
- Turbo and Nano generation requires a reference clip for voice cloning, per the README examples.
- If the reference clip's language differs from the language tag, output may inherit the reference accent unless cfg_weight is set to 0.
- Every generated file carries a Perth watermark, which the README presents as built in.
Is Chatterbox growing? Usage data
In the week of Oct 25–31, 2026, Chatterbox ranked #80 on the Devtoolwatch Board and #3 in Voice and realtime, with a momentum score of 68 out of 100. Its main package had 73.2k downloads that week, +44% against its own 4-week average.
Show the weekly numbers
| Week of | New stars |
|---|---|
| Aug 9, 2026 | 99 |
| Aug 16, 2026 | 121 |
| Aug 23, 2026 | 102 |
| Aug 30, 2026 | 97 |
| Sep 6, 2026 | 113 |
| Sep 13, 2026 | 105 |
| Sep 20, 2026 | 96 |
| Sep 27, 2026 | 104 |
| Oct 4, 2026 | 185 |
| Oct 11, 2026 | 219 |
| Oct 18, 2026 | 185 |
| Oct 25, 2026 | 285 |
On Hacker News
- Chatterbox TTS, 670 points and 188 comments (Jun 11, 2025)
- Chatterbox, Resemble AI's production-grade open source TTS model, 6 points and 1 comments (May 31, 2025)
- Chatterbox – open-source TTS model, 5 points and 0 comments (May 28, 2025)
The Board ranks tools by four signals each week: new stars, new stars for the repo's size, added downloads and download growth beyond the market. How the ranking works.
Chatterbox alternatives
Other voice and realtime AI tools on the Board in the week of Oct 25–31, 2026, by momentum:
| Rank | Tool | Stars | Downloads/wk |
|---|---|---|---|
| #1 | faster-whisperfaster-whisper is a Python reimplementation of OpenAI's Whisper speech-to-text model built on the CTranslate2 inference engine for… | 26.4k | 4.4M |
| #2 | KokoroKokoro is a Python inference library for Kokoro-82M, an open-weight text-to-speech model with 82 million parameters. | 9.8k | 150.7k |
| #4 | VoiceMemInfrastructure for the next generation of voice agents, designed to provide universal memory. It is divided into a left brain and a… | 2.8k | – |
| #5 | PipecatPipecat is an open-source Python framework for building real-time voice and multimodal conversational agents, maintained by Daily and… | 17k | 348.5k |
| #6 | WhisperXWhisperX is a Python speech recognition tool built on Whisper that adds word-level timestamps and speaker diarization. | 25k | 158.6k |
| #7 | LiveKit AgentsLiveKit Agents is an open-source Python framework for building realtime voice, video, and multimodal AI agents that join LiveKit rooms… | 15k | 2.4M |
Recent Chatterbox releases
- v0.1.2, Jun 13, 2025
Chatterbox FAQ
Is Chatterbox free and open source?
Yes. Chatterbox's source code is on GitHub (resemble-ai/chatterbox) under the MIT license, so you can use it for free.
How do I install Chatterbox?
Run pip install chatterbox-tts. The chatterbox-tts package needs Python 3.10 or newer. The install section above has the other options from the project's README.
Is Chatterbox still maintained?
Activity has slowed: the resemble-ai/chatterbox repository was last updated on Jul 21, 2026, and the latest release, chatterbox-tts 0.1.7 on PyPI, came out on Mar 26, 2026.
How popular is Chatterbox?
Chatterbox has 27.5k GitHub stars as of Nov 1, 2026. It gained 285 new stars in the week of Oct 25–31, 2026, against a 12-week average of 143 a week. Its main package had 73.2k downloads that week, +44% against its own 4-week average. That made it #3 among voice and realtime AI tools on the Devtoolwatch Board.
Chatterbox vs Kokoro: how do they compare?
Chatterbox and Kokoro are both tracked by Devtoolwatch as voice and realtime AI tools. Chatterbox has 27.5k GitHub stars and Kokoro has 9.8k. In the week of Oct 25–31, 2026, Chatterbox's main package had 73.2k downloads and Kokoro's had 150.7k. Kokoro is a Python inference library for Kokoro-82M, an open-weight text-to-speech model with 82 million parameters. Kokoro review.
What are the best alternatives to Chatterbox?
Other voice and realtime AI tools on the Devtoolwatch Board in the week of Oct 25–31, 2026, by momentum: faster-whisper (#1), Kokoro (#2), VoiceMem (#4), Pipecat (#5), WhisperX (#6). The alternatives table above has their numbers.
How do I use Chatterbox TTS locally?
Chatterbox runs locally as a Python package. Install chatterbox-tts, load a model such as ChatterboxTTS.from_pretrained(device="cuda") or with "cpu" or "mps", call generate with your text, and save the result with torchaudio.
How do I use Chatterbox Turbo?
Chatterbox Turbo is loaded with ChatterboxTurboTTS.from_pretrained(device="cuda") from chatterbox.tts_turbo. Pass text, optionally with tags like [chuckle], and a reference clip via audio_prompt_path to generate.
What is the difference between Chatterbox and Chatterbox Turbo?
Chatterbox Turbo is a 350M-parameter English model for low-latency voice agents, with a one-step decoder and paralinguistic tags. The original Chatterbox is a 500M English model with CFG and exaggeration tuning for general zero-shot TTS with creative controls.
Is Chatterbox open source?
Chatterbox is open source. The repository and homepage describe it as an open-source TTS model family from Resemble AI, published on GitHub and Hugging Face.
Which languages does Chatterbox support?
Chatterbox Multilingual supports 23 languages, including Arabic, German, Spanish, French, Hindi, Japanese, Korean, Russian, and Chinese. Dedicated Single Language Pack models cover Chinese, Latam and Spain Spanish, Brazilian and Portugal Portuguese, and Hindi.
Can I detect whether audio was made with Chatterbox?
Chatterbox embeds a Perth watermark in every generated file. The README provides a script using perth.PerthImplicitWatermarker().get_watermark, which returns 1.0 for watermarked audio and 0.0 otherwise.
Sources
- resemble-ai/chatterbox on GitHub
- Chatterbox README
- chatterbox-tts on PyPI
- Chatterbox website
- Release v0.1.2 (Jun 13, 2025)
- Hacker News: Chatterbox TTS
- Hacker News: Chatterbox, Resemble AI's production-grade open source TTS model
- Hacker News: Chatterbox – open-source TTS model
- Devtoolwatch Board, the week of Oct 25–31, 2026
Chatterbox README badge
Maintainers can show the tool's weekly Board rank in their README:
[](https://index.devtoolwatch.com/tools/chatterbox/)How this page is made: Devtoolwatch writes each review from the project's README, releases, package metadata and website, with an AI model, then checks every claim against the source it cites and drops what the sources don't support. The numbers come from Devtoolwatch's daily GitHub snapshots, GitHub's stargazer history, and the npm and PyPI download APIs. Spotted an error? Email [email protected].