# Create an audio overview (https://docs.commonplot.co/en/user/audio-overview/)

An audio overview turns a notebook's sources into a conversation between two hosts, so you can review a topic while commuting or doing something else.

## Create an audio overview [#create-an-audio-overview]

1. Open a notebook and switch the AI rail to **Studio**.
2. Choose **Audio overview**.
3. Under **Sources**, select the resources and documents to use. Every source starts selected, and there is no limit on how many.
4. Choose settings:
   - **Language**: English or Chinese
   - **Target length**: Brief (about 1–2 min), Standard (about 2–3 min), or In depth (about 4–5 min)
   - **Focus (optional)**: what the two hosts should discuss
5. Choose **Generate audio**.

![Audio overview settings: sources, target length, language, and focus](https://docs.commonplot.co/images/en/audio-overview-dialog.png/)

You can also ask in AI chat: "Make an audio overview of this notebook."

## While it generates [#while-it-generates]

Generation moves through **Reading the sources and writing**, **Producing audio**, and **Saving the result**. The AI decides how closely to read each source for your focus, so more sources take longer and use more AI usage. Leaving the page doesn't cancel it. Up to three generations can run at once.

If a generation fails, **Retry** runs it again with the same sources and settings, reading the sources as they are now.

## Listen and manage [#listen-and-manage]

Finished audio overviews appear in **Studio** and in the notebook's resources rail. The result page has:

- A player
- The script
- The sources used, with links to open them

**Delete audio overview** deletes only the audio, not its sources. Audio files count toward your [storage](https://docs.commonplot.co/en/user/files/).

## AI usage [#ai-usage]

Generating an audio overview uses AI usage. If you run low, **Manage AI allowance** shows your plan options. See [Plans and AI usage](https://docs.commonplot.co/en/user/plans-and-ai-usage/).