How it works
- Choose Cloud for fast results or On-device to run the model in your browser.
- Add a reference sample: 10 to 20 seconds of clear speech works best.
- Type the text you want spoken and generate.
Cloud or on-device: which should I use?
Cloud cloning is fast and includes a free preview of about one minute of audio with no account. Paid plans include more cloning time.
On-device cloning runs the Chatterbox model in your browser, so your sample never leaves your computer. It's much slower: on a CPU, expect around a minute per 10 words.
How do I get a good clone?
Use a quiet room, one speaker, and natural sentences at a normal pace. Background music, echo and overlapping voices all make the result worse.
Whose voice can I clone?
Your own, or a voice you have clear permission to use. Don't clone someone's voice without their consent, and don't use a cloned voice to impersonate or mislead anyone.
Frequently asked questions
Is voice cloning free?
You can try a cloud clone of about one minute with no account, and the on-device option runs free in your browser. Paid plans include more cloud cloning time.
How much audio do I need?
About 10 to 20 seconds of clear speech from one speaker.
Is my voice sample private?
With on-device cloning, the sample is processed in your browser. Cloud cloning sends the sample to the cloning service to generate the audio.
Can I clone someone else's voice?
Only with their permission. Use voice cloning for your own voice or voices you're authorized to use.