Skip to content

AI AudioEarly access

The parts

And around it

Describe. Generate. Keep.

Takes, with waveforms.

Who it is for

What people do with it

Four ways to work.

Start here

Look things up

Documentation

From the first take to Media.

Read

Know us

Read, ask, follow along

Notes on voices, beds and honest previews.

Voice-overs

How to make an AI Audio voice-over: text, language English, German or Dutch, a voice, a speed from 0.5 to 2 and an emotion, one to four takes.

Updated 18 October 2026 · View as Markdown

Text. What to say.
Language. English, German or Dutch.
Voice. The Preview voices are Ava, Ben, Noor, Max and Robo. A real model has its own voices.
Speed. From 0.5 to 2.
Emotion. Neutral, warm, cheerful, calm, serious or excited.
Takes. One to four.

The cost is 1 credit per 500 characters, times the takes. The Preview voice is a robot voice: pitch from the voice, contour from the emotion and rate from the speed.