Skip to content

Narration

How to turn on AI narration, review the scripts, generate the audio, and how narration tokens are counted.

AI narration generates a voiceover from your course text. Turn it on, pick a voice, review the scripts, then generate the audio when you publish.

AI narration automatically generates a high-quality voiceover for your course content, so learners can listen as well as read. You turn it on, choose a voice, review the narration script for each screen, and the audio is generated when you publish. Generating narration uses tokens from your narration allowance.

Turn on narration

  1. Open the course and select the Narration icon at the top of the screen.
  2. Toggle on Use Narration.
  3. Choose your voice, and adjust the autoplay and progress settings. The narration language is set automatically from the language your course is written in, so there's nothing to select there.
  4. Select Save changes. If you don't save, narration won't stay on and you won't be able to generate audio.

Review the narration scripts

By default, narration reads the on-screen text. If you want the narration to differ from what's on screen:

  1. Open a screen and select the Narration settings icon.
  2. Toggle off Sync narration with screen.
  3. Edit the narration script for that screen, and repeat for any others.

Reviewing scripts before generating avoids using tokens on audio you'll redo.

Generate the narration

Narration is generated when you publish:

  1. Select Publish, then toggle on Generate narrations.
  2. The dialog shows how many narrations will be generated and your token balance.
  3. Select Generate narrations and publish.

Generate narration only once your course is finalised, to make the best use of your tokens.

Choose and preview a voice

In Course settings, then Narration, choose from the available voices. The narration language matches your course automatically, so you're choosing a voice within that language. Select the speaker icon next to a voice to hear a sample. Changing the voice after generating means publishing again to regenerate the audio.

Autoplay and listening settings

  • Autoplay: narration starts automatically when a learner lands on a screen (except the first screen, which most browsers block until the learner interacts).
  • Require learners to listen in full: learners can't move on until the narration finishes, which keeps them on each screen for its full audio.

The learner experience

When narration is on, learners see a speech-bubble icon in the bottom-right. They can pause and resume narration, and adjust playback speed while it plays.

Narration tokens

One narration token equals one character of generated narration, so a 1,200 character script uses 1,200 tokens. Standard and premium voices draw from the same pool, and a translated course narrates each language separately, so it uses roughly twice the tokens of the same course in one language.

The allowance is hard capped. When you reach 100%, narration generation stops until the allowance resets at the start of the next month, and there's no overage charge. Everything else keeps working: courses stay live, learners keep taking them, and you can keep building, editing and publishing without narration.

That's why you generate audio at publish time rather than on every edit. The publish dialog shows how many narrations will be generated and your current token balance before you commit, and your current usage is on the Usage limits page in Settings.

Working with the AI agent

If you're building with the AI agent, you can ask it which narration voices are available and how much of your narration allowance is left, so you can check before committing to a generation.

Frequently asked questions

Can I change the narration language?

No. The language is detected automatically from the language your course is written in. You choose a voice within that language rather than setting the language yourself.

Do edits use tokens?

Tokens are only used when you generate or regenerate audio by publishing. Editing scripts or changing settings like autoplay won't use tokens until you publish.

Can I edit a narration script after generating?

Yes. Edit the script any time, then publish again to regenerate that screen's narration, which uses additional tokens.

What happens if I run out of tokens mid-course?

Narration generation stops for the rest of the month and starts again at the reset. Your course still publishes, and everything except the narration audio generates as normal.