Clone your voice — free, unlimited, on your device
Record ten seconds of yourself reading, then type anything and hear it in your own voice, in English, French, German, Spanish, Portuguese or Italian. The model runs in your browser: no account, no length limit, and your files are never uploaded.
Your voice
Read this aloud — about 10 seconds
The sun comes up over the city. I make a coffee, open the window and listen to the street below.
Drop an audio file — 5 to 30 seconds of one person speaking — or record yourself:
Your text
Speech
How to clone your voice
-
Record yourself, or drop a recording
Read the sentence on screen for about ten seconds, or drop a clip of you speaking — 5 to 30 seconds of one person, with little background noise.
-
Confirm it is your voice
Tick the box: it is your own voice, or the speaker has explicitly agreed. The tool will not use a sample until you do.
-
Type, generate, download
Type or paste any text and download a WAV in your voice. Long text is split at sentence ends and spoken in turn.
Why clone your voice here
Your files are never uploaded
The model learns your voice from the recording on your device, in your browser. The recording is not sent to a server at any point — you can check that in the Network tab while it works.
No length limit
Paste a page or a chapter. There is no character cap and no daily quota, because the work happens on your computer, not on a server.
Six languages
Your voice can speak English, French, German, Spanish, Portuguese and Italian. It sounds most like you in the language you recorded in.
The model behind it
Pocket TTS by Kyutai, a 100-million-parameter voice model (CC-BY 4.0) small enough to run on a laptop's processor. The first use downloads about 140 MB.
Kept only if you ask
Tick “keep” and your recording is stored in this browser only, so next time your voice is ready. One click deletes it.
Questions
- Is it really free and unlimited?
- Yes. There is no account, no character limit and no daily quota. The model runs on your own device, so a longer text costs us nothing — it just takes your computer a little longer.
- Is my voice recording uploaded?
- No. The recording is turned into a voice inside your browser and is never sent anywhere. To check, open your browser's developer tools, switch to the Network tab and clone your voice with it open: you will see the model download once (about 140 MB) and no request carrying your recording. This page loads analytics like most websites do, which is why the claim is precise: your files are never uploaded.
- Can I clone someone else's voice?
- Only with their explicit permission. Cloning a voice without consent is prohibited by the model's terms and, used to impersonate or deceive, is illegal in many countries. The tool asks you to confirm before it uses any sample.
- How much audio do I need?
- 5 to 30 seconds of one person speaking clearly. Reading the sentence on screen for about ten seconds is enough. A quiet room helps more than a better microphone.
- Which languages can my voice speak?
- English, French, German, Spanish, Portuguese and Italian. Your cloned voice sounds most like you in the language you recorded in; in the others it keeps your tone with that language's accent.
- Which model does it use?
- Pocket TTS by Kyutai, released under CC-BY 4.0, converted to run in the browser on WebAssembly. It needs no graphics card.
- What do I get?
- A 48 kHz mono WAV file in your voice, ready to play or drop into a video editor. (The model itself renders at 24 kHz; the download is resampled to the 48 kHz video timelines use.)
Is it legal to clone a voice, and how do ten seconds become one? How voice cloning works — what the model does with your recording, what US and EU law say about cloning other people, and why most services keep your voice.