Clone a voice

Record or upload 10–20 seconds of someone speaking, and VoxStudio will speak any text in that voice.

Only clone voices you have permission to use. VoxStudio asks you to confirm consent and saves that statement with the voice. Everything a cloned voice says carries an inaudible watermark identifying it as AI-generated.

Before you start

Voice cloning uses the Chatterbox model. If it isn't installed yet, the Clone page shows a Download card (3.2 GB plus a 1.3 GB engine, one time only). You can keep using VoxStudio while it downloads.

1. Capture a sample

Record:

  1. Click Record. The first time, allow microphone access.
  2. Read the passage on screen at your normal pace. The level meter should sit mostly in the green.
  3. Click Stop once you've finished: 10–20 seconds is ideal.

Or upload: drop a WAV, MP3 or M4A file onto the upload area, or click it to choose one. It should be 10–60 seconds of one person speaking, with no music.

VoxStudio checks the sample right away:

Check What it means
✅ Great recording Clear, a good length, no distortion
⚠️ A bit short It works, but 10–20 s matches better
⚠️ A little quiet / Some distortion Move closer or further away and record again
⛔ Too short / Too long / Very quiet Record again before continuing

Press ▶ to listen, or Redo to start over.

2. Name it and confirm consent

Give the voice a name, tick the consent statement, and click Create voice.

3. Try it out

Type a test sentence and click Hear it. The first sentence takes longest while the engine warms up. When you're happy, click Use in Studio to start writing with your new voice.

Your cloned voices appear at the bottom of the Clone page (click one to hear its original recording) and under Your voices in Studio.

Tips for a great clone