Documentación

Voice cloning lets you create an AI version of your own voice from a few minutes of recordings. Once created, your cloned voice appears in the voice selector alongside the stock voices, so you can generate text-to-speech audio that sounds like you ? without recording a word.

Voice cloning is available on paid plans. The number of voice clones you can create depends on your plan - check My Voices to see your current count and limit.

Creating a voice clone

Go to My Voices > Create a clone and fill in:

1. Name and description - how the voice appears in your voice selector.
2. Gender, age, and accent - these are notes to help identify your voice.
3. Voice samples - pick 1 to 4 audio files from your audio library, up to about 3 minutes of audio in total. Each file must be under 11 MB.

Click Create and watch the progress - when processing finishes, you're taken to your new voice's page, where you can play a sample of the cloned voice.

Getting good results

• Use clean recordings of just your voice - no music, background noise, or other speakers.
• Natural, conversational speech clones better than reading in a monotone.
• If the clone doesn't sound right, delete it and try again with a different set of samples. If a sample set can't be processed, you'll be prompted to try different recordings.

Consent and fair use

You must have permission to use the audio you upload. Cloning someone's voice without their consent is not allowed, and unauthorised use of audio files may result in your account being suspended. And remember: our usage policy requires you to disclose the use of AI-generated speech wherever you share your videos.

Managing your voices

My Voices lists all your cloned voices with their status. Open a voice to edit its name, description, gender, age, or accent, preview it, or delete it. Deleting a voice frees up a slot for a new clone.