Audio

• Voice-over — natural speech from a script of up to 5,000 characters.
• Music — a quick ~30-second clip or a full song of up to about 3 minutes, with or without lyrics.
• Sound Effects — whooshes, ambiences, footsteps and more, up to 30 seconds, optionally looping.
• Lip Sync and Talking Video — make a face say a line.

Audio tools need Sketch or above.

By the length of your script: longer scripts cost more. The Generate button updates the price as you type, so you always see the cost before you run.

ElevenLabs Multilingual v2 speaks 29 languages and ElevenLabs v3 speaks 74. Filter the voice library by language, gender, accent and use.

Voice Clone is rolling out to Canvas and above. You record or upload a clean 10–90 second sample and confirm that it is your voice, or that you have the speaker’s written permission.

1. Choose who talks: an image, a video or a Cast member.
2. Choose what they say: pick audio from your Library, record with your microphone, or type a line to turn into a voice-over first.
3. Confirm you have consent to use the person’s likeness, then generate.

Lip Sync is billed on the length of the audio, rounded up to the next second.

The same rule applies to all outputs: commercial use needs Canvas, Pro, Atelier or a team plan at the time you create it.