Skip to content

How to generate and edit audio descriptions

This guide covers every option on the video-creation form and every action available in the description editor afterward. If you haven’t created a video project before, start with Describe your first video instead — this guide assumes you already know the basics.

Select New Video Project in the sidebar. Provide a source and name under Video Source, and a language under AI Settings — see Describe your first video if you need a walkthrough of those basics — then set any of the following before you select Create. All are optional except the source, name, and language.

The Estimated Cost panel above the button updates as you change these, so you can price a choice before making it.

If your source is a YouTube URL, note that export later is limited to SRT only — see Export a video. For file size and length limits by plan, see Project types and file limits.

Choose the narration template that matches the content — for example, a documentary or cinematic style versus a straightforward accessibility style. This changes the tone and phrasing of every generated description, not just individual words.

If you want our tool to know things it can’t infer from the video alone — character names, brand terms, pronunciation notes, or content to avoid describing — put them here. The field accepts up to 10,000 characters.

If you have several recurring characters or terms, attach a glossary instead of restating them here every time — see the next section.

If your video has recurring characters, people, or terms that should be described consistently, attach a glossary to the project so descriptions use the canonical names you defined. Pick one from the Glossary list, or create one first via Manage glossaries — see Build a glossary of characters and terms.

If you want a consistent house style — tone, length, terminology — reused across many projects, attach a Description profile instead of retyping instructions each time. Manage profiles opens the profile list in a new tab.

Choose how much detail each description includes:

  • Concise — the shortest descriptions that still convey the essential action.
  • Detailed — balances completeness and pacing.
  • Comprehensive — the most thorough descriptions, useful for content where visual detail carries a lot of meaning.
  • Standard — a single processing pass.
  • Enhanced — re-checks each description with an additional review pass before finalizing it. Produces more accurate output, at a higher credit cost. If accuracy matters more than cost for this project, choose Enhanced.

For exact credit costs of Standard versus Enhanced, see Plans and credit costs.

If you also want captions generated alongside the audio descriptions, turn on the Captions toggle. It adds error correction and sound labels, and costs extra credits — the Estimated Cost panel shows the addition under Subtitles. See Turn on and refine captions for what that adds.

Select Create. Processing then runs in the background — you can leave the page or close the tab, and the project’s status moves from Pending to Processing to Completed. Processing time depends on video length and the quality you chose.

If a run takes longer than ten minutes, we email you at your account address when it finishes, with a link back to the project. There’s nothing to enable and no address to enter. Runs under ten minutes send no email, since you’re most likely still on the page. On an organization project the email goes to whoever started the run, not to the organization’s owner.

Review and edit descriptions in the editor

Section titled “Review and edit descriptions in the editor”

Open a completed project to reach the description editor. It shows a timeline of AI-generated descriptions aligned to the video.

If a description is technically correct but too sparse, use Elaborate (or ask our tool) on that description to get more detail without discarding the original. This is useful when you want more context without a full rewrite.

If a description needs to change — wrong detail, wrong tone, missed action — select Rewrite (touch-up) to have our tool regenerate that specific description. This replaces the description rather than adding to it.

If you want to change the exact wording yourself instead of asking our tool to regenerate it, edit the description text directly in the editor.

Both elaborate and rewrite are billed separately from the initial generation — see Plans and credit costs for current rates.

If you want to check a detail against the video without editing anything, use the ask-a-question feature to query our tool about the project’s content.

Once you’re happy with the descriptions themselves:

  • Tune the narration — choose the voice and speed, decide how narration sits under the original audio, and adjust the timing of individual lines.
  • Export a video — produce the final MP3, MP4, or SRT file.
  • Share a video — publish a link where each viewer chooses how descriptions reach them.