Skip to content
Sign in Add to Chrome

Voiceovers

Give every step a voice.

Write a short script for each step, then pick an AI voice, record your own or upload a file. Captions are timed to the words, and every language you translate gets its own narration.

Try Pro free for 14 days See pricing
  • Free (not included)
  • Pro (included)
  • Team (included)

AI voice comes with 60,000 characters per creator a month on Pro (about an hour of speech) and 120,000 on Team. Trials get 10,000. Recording and uploading your own audio don’t count.

Narration

An AI voice, your voice or a file.

Open Voiceover from the step list. Each guided step and each clickable screen gets its own script, separate from the tooltip and usually shorter. Then choose how it’s spoken.

  • Filter AI voices by gender and accent, play samples, and set the speed (0.7× to 1.2×) and stability
  • Record with your microphone in the editor. Noise is suppressed, silence trimmed and the level evened out
  • Upload audio as MP3, M4A, WAV, WebM or Ogg, up to 3 minutes a clip
  • Start from Copy from tooltips, or let Write with AI draft the scripts (AI credits)

Captions and playback

Captions that follow the words.

AI clips get captions timed from the voice itself. Recordings and uploads are transcribed, and the transcript becomes the script and the captions. Fix any line with Edit captions.

  • Captions on by default, or leave viewers to turn them on with CC
  • Advance after narration moves a guided step on when its clip ends
  • In a clickable demo, a screen’s narration plays when the screen opens
  • Sound comes on with the viewer’s first click in the demo, never before

Languages and names

Narrate in every language you ship.

Scripts are translated with the rest of the demo. On the Languages tab, generate each language’s narration in the same voice, and its captions come with it.

  • A language without narration for a step plays that step silently, never in the wrong language
  • Add “Say it as” to a glossary term and every AI clip pronounces it that way
  • Variables are spoken with their default value, so one clip serves every viewer and no viewer data goes to the voice
  • Editing a step marks only its clip Out of date, so you remake one clip, not the whole demo

Allowance and analytics

Know what you have, and who listened.

AI voice is measured in characters of script. Each Pro creator seat adds 60,000 characters a month to the workspace, about an hour of speech, and each Team seat adds 120,000. A 10-step demo is about 1,200 characters.

  • Before you generate, the editor shows how many characters it will use and how many are left
  • Recording and uploading never count, and Generate skips clips that haven’t changed
  • Analytics show sessions with narration, how many turned sound and captions on, and how much of each step was heard

How it works

  1. 01

    Write the scripts

    Copy them from the tooltips, draft them with AI or write your own.

  2. 02

    Make the clips

    Generate them with an AI voice, record yourself or upload audio.

  3. 03

    Publish

    Narration and captions go live with the next publish.

Use cases

Teams that use it

FAQ

Questions, answered.

Which plans include voiceovers?

Pro and Team. During the 14-day trial, AI voice is limited to 10,000 characters a month (about 10 minutes of speech), and recording or uploading your own narration doesn’t count against it. Demos on the Free plan play without narration. See pricing.

Is the allowance per person?

It’s per creator seat, shared by the workspace. A Pro workspace with three creators gets 180,000 AI voice characters a month to use on any demo. The allowance resets on the 1st of each month (UTC).

What happens when we run out of characters?

AI voice generation stops until the 1st of next month. On a paid plan, adding a creator seat raises the allowance. Clips you already made keep playing, and you can still record or upload narration.

Can I clone my own voice?

No, AtriumKit doesn’t clone voices. To use your own voice, record it in the editor or upload an audio file.

Can narration say the viewer’s name?

No. Variables in a script are spoken with their default value, so one clip serves everyone and no viewer’s details are sent to make audio. Use personalization for names and companies in the on-screen copy.

Works well with

Your next demo is one click-through away.

Add to Chrome Open the app