Resemble AI (Chatterbox)
Resemble AI Chatterbox: open-source speech and voice cloning
Chatterbox is Resemble AI’s open-source text-to-speech family for local speech generation and voice cloning. It includes multilingual models, English Turbo and the smaller Nano model for on-device use.
FreeAccount optional
- Access
- Free
- Available on
- Desktop, API
Speech models you can run and adapt yourself
Chatterbox offers a self-hosted route to AI narration and custom voices. Resemble AI keeps its generative voice models available as open-source research, while its commercial business now focuses on deepfake detection. In August 2026, the company announced that it would continue supporting existing voice customers but stop selling voice AI to new ones. For a new speech project, the relevant offering is therefore Chatterbox’s downloadable model family rather than a new hosted Resemble voice subscription.
The family includes general multilingual speech, English Turbo and English Nano. Turbo targets applications with tighter compute requirements; Nano reduces the model further for CPU and on-device use. Original Chatterbox provides expressive exaggeration and guidance controls, while Turbo and Nano support vocal tags such as laughter. Their controls are different: the Turbo implementation ignores the original model’s exaggeration and CFG settings. Reference audio supplies the voice identity, and the local examples save generated audio as WAV.
The MIT software licence allows modification and commercial deployment while retaining its notice. Local hosting also gives a team responsibility for infrastructure, integration and updates; it is a software model family rather than a finished browser production studio. Default generation includes Resemble’s PerTh watermark. Language quality varies: the Multilingual V3 release specifically cautions against commercial Korean and Vietnamese use and reports limitations in Czech, Japanese and Finnish. A software licence does not provide permission to clone someone else’s voice.
Best for
- Developers building self-hosted speech applications
- Teams needing local voice generation
- Researchers adapting open-source voice models
Limitations
Resemble no longer accepts new paid voice AI customers. Chatterbox requires technical setup and self-managed compute. Turbo/Nano controls differ from the original model. The Multilingual V3 release warns against commercial Korean/Vietnamese output. Voice and input-content rights are separate from the MIT software licence.
Chatterbox speech-generation features
Local speech generation
Run the open-source models on your own infrastructure instead of sending each generation to a hosted voice service.
Multilingual model family
General multilingual models and separate language/dialect fine-tunes serve localisation projects.
Reference-based custom voices
Generate speech using a supplied voice recording; Turbo’s example starts with a 10-second clip.
English Turbo and Nano
Choose English models aimed at reduced compute needs, including a smaller Nano model for CPU/on-device deployment.
Model-specific expression
Original-model exaggeration controls and Turbo/Nano vocal tags provide different ways to shape speech.
Open-source licensing
The MIT software licence permits adaptation and commercial use with the required notice; default audio includes a PerTh watermark.
Resemble AI (Chatterbox) pricing
Free to use. Custom pricing. Visit Resemble AI (Chatterbox) for full plan details and current offers.
Technical specifications
Voice generation
| Feature | Resemble AI (Chatterbox) |
|---|---|
| Speech languages | The repository lists 23 general multilingual languages and six specialised language/dialect models. Turbo and Nano are English models. Coverage and quality depend on the chosen checkpoint.Multilingual versus English model selection |
| Voice and pronunciation controls | Original Chatterbox offers exaggeration and guidance controls. Turbo/Nano use native paralinguistic tags such as laughter; their code ignores exaggeration and CFG adjustments.Controls differ between model architectures |
| Voice cloning | Reference-based cloning runs locally. Turbo requires a clip longer than 5 seconds and uses a 10-second decoder reference window; the repository example uses a 10-second clip.Self-hosted; speaker permission and audio rights required |
| Audio exports and streaming | The Python examples save generated speech as WAV. Default generation embeds a PerTh watermark. There is no included managed export or hosted service for new paid voice customers.Local output and self-managed deployment |
| Commercial-use terms | The MIT licence permits commercial use and modification of the software with its licence notice. Rights to a speaker’s voice and other input content remain separate.MIT; input rights remain the user’s responsibility |
| API access | SupportedSelf-hosted integration; no new paid hosted voice accounts |
Chatterbox alternatives
Hosted speech platforms provide managed generation and account-based usage.
Cartesia
Managed Sonic speech APIs with streaming, multilingual voices and paid cloning.
Compare with Resemble AI (Chatterbox)ElevenLabs
A hosted speech platform with a browser studio, multilingual generation and voice cloning.
Compare with Resemble AI (Chatterbox)Murf
A browser voiceover editor for script-based narration and production exports.
Compare with Resemble AI (Chatterbox)Resemble AI (Chatterbox) FAQs
Can new customers buy Resemble AI voice generation?
Resemble’s August 2026 announcement says it no longer sells voice AI to new customers, while continuing support for existing voice customers. Its Chatterbox models remain open source and free.
Is Chatterbox free for commercial projects?
The software is distributed under MIT, which permits commercial use and modification with its licence notice. Hardware, hosting and maintenance costs are separate, as are permissions for voice recordings and other input content.
Does Chatterbox run locally?
Yes. The public Python package and examples support self-hosted generation. Nano targets CPU and on-device use; hardware requirements and available controls depend on the selected model.
Are Chatterbox Turbo and Multilingual the same model?
No. Turbo and Nano are English models with native vocal tags. Multilingual covers other languages, and the original model has different expression controls. Some languages have additional release-specific quality limitations.
Is this your product? Claim this page to update your listing.