NotebookLM captions and accessibility: what WCAG asks for

The gap

NotebookLM renders a Video Overview with play, speed, and fullscreen controls, and it renders an Audio Overview with the same set. Neither ships a caption toggle and neither exposes a timed text track, so the narration is unavailable to anyone who cannot hear it.

What WCAG 2.2 asks for

• SC 1.2.2 Captions (Prerecorded), Level A: synchronised captions for all prerecorded audio in video.
• SC 1.2.1 Audio only and video only, Level A: a transcript for audio only content such as an Audio Overview.
• SC 1.2.3 Audio Description or Media Alternative, Level A: a text alternative that conveys what the visuals show.
• SC 1.2.5 Audio Description, Level AA: actual audio description for prerecorded video.
• SC 1.4.3 Contrast (Minimum), Level AA: caption text needs at least 4.5 to 1 against its background.
• SC 1.4.4 Resize Text, Level AA: text scales to 200 percent without loss of content.

Captions and transcripts cover the first two. Audio description is a different job, because it describes visual information that is not spoken, and no automatic tool does it well yet.

Where automation helps and where it does not

Generating captions from the narration closes the Level A caption and transcript gaps in minutes. The captions are machine generated, which means they inherit whatever the model misheard. For personal use that is usually good enough, and the inline editor lets you correct a cue in place. For anything published, assessment, or legally required, a human pass over the transcript is still the standard.

Practical checklist for an accessible overview

1. Generate captions before you shareCaptions for Video Overviews, and a caption dock plus transcript for Audio Overviews.
2. Read the transcript onceFix the words the model got wrong, especially proper nouns.
3. Check contrast and sizeCaption styling in LM Subs lets you raise size and background opacity until the text is legible on bright slides.
4. Export a transcript alongside the videoA Markdown or TXT export satisfies the media alternative requirement and helps anyone who would rather read.
5. Describe the visuals separatelyIf slides carry meaning the narration never states, write that description by hand.

Beyond captions

Captions help far more than the deaf and hard of hearing audience they were written for. Non native speakers, people working in a quiet office, and anyone who reads faster than speech all benefit from a transcript attached to the timeline.

Frequently asked questions