WebGPU flagship
Chatterbox: specs & requirements
Chatterbox is Resemble AI's 0.5-billion-parameter model — the highest-quality open TTS that runs in a browser, with zero-shot voice cloning from a few seconds of microphone audio. It demands a real GPU via WebGPU and a ~1.5 GB one-time download, but the output rivals paid cloud services.
Chatterbox support is being wired into the tool. Its specs and hardware verdicts are below — meanwhile, the live models are ready to use, and the compatibility check already tells you whether your device could run Chatterbox.
Specifications
| Developer | Resemble AI |
|---|---|
| Parameters | 500 million |
| Download size | 1550 MB (q4 mixed) |
| Output | 24.0 kHz mono |
| Voices | 1 |
| Languages | English (US) |
| Voice cloning | Yes |
| Streaming | No |
| License | MIT |
| Speed | Near real-time on a desktop WebGPU GPU; impractical on CPU. |
| Runtime | WebGPU required |
Source: onnx-community/chatterbox-ONNX
Strengths
- +Best quality of any browser-runnable open model
- +Clone any voice from a 5–10 second sample, entirely on-device
- +Plain MIT license
Limitations
- −Requires WebGPU and ~8 GB RAM — desktop-class hardware
- −1.5 GB download before first use
Chatterbox — questions
Is Chatterbox free to use?
Yes. The weights are MIT and they run on your device.
What hardware does Chatterbox need?
A WebGPU-capable browser and GPU (Chrome/Edge, Safari 26+, Firefox 141+) plus about 8 GB of RAM. It will not run on CPU at usable speed.
Is my text private when using Chatterbox here?
Yes. The model executes inside your browser tab — the text you synthesize is never transmitted to a server.
Can I use Chatterbox output commercially?
Review the current model, voice, and source-dataset terms before commercial use. The catalog lists MIT for this model.