In short. A descriptive video transcript provides speech, relevant sounds and essential visual information in a readable text alternative.
Definition: A video transcript is a text version of the speech and meaningful sounds in a video. A descriptive transcript also records the visual information needed to understand the content, making it available through text.
Key facts
- The W3C Web Accessibility Initiative's transcript guidance identifies descriptive transcripts as necessary for people who are both Deaf and blind.
- W3C's caption guidance explains that captions display audio information in the media player, synchronized with the sound.
- W3C's Understanding Success Criterion 1.2.8 describes a text alternative that preserves the sequence of dialogue and visual descriptions.
- As of October 2026, Google's AI features documentation recommends making important content available in textual form for AI Overviews and AI Mode.
What information belongs in the transcript?
W3C's transcription guidance calls for speech and relevant non-speech sounds, such as a baby crying or approaching horse hoofs. It recommends identifying speakers where relevant and including background music when it contributes to understanding. Descriptions of music use objective language about its mood.
Visual information requires a separate review. The transcript guidance gives examples of text that captions may omit, including the video's title and a speaker's name and job title. A descriptive transcript adds those details when needed for understanding.
The same guidance provides this visual description in its example transcript:
A man sitting at a desk starts watching a video on his computer.
W3C's explanation of Success Criterion 1.2.8, part of the Web Content Accessibility Guidelines (WCAG), extends the requirement to interaction. If a video asks the viewer to respond, the text alternative provides links or another means of performing the equivalent action. Its descriptions follow the order of events in the video.
How is a transcript prepared and checked?
W3C's transcript guidance describes reusing a caption file: export plain text where the editing tool supports it, or remove the timestamps. Caption lines can be combined into logical paragraphs. Timestamps remain optional when they help readers use the transcript.
Accuracy needs a separate check. The caption guidance says automatic captions meet user needs and accessibility requirements only when confirmed fully accurate. An automatic transcript therefore needs comparison with the recording before publication.
A production script also needs comparison with the final media. W3C's explanation of the media alternative criterion accepts a screenplay as an alternative only when corrected to represent the edited result accurately. Reviewing the finished video also establishes the sequence in which visual descriptions belong.
Where does the text appear on a website?
W3C's transcript guidance says most web transcripts use HTML. For media hosted on the website, it generally recommends placing the transcript on the same page. The text itself, or a link to it, can appear directly beneath the video. Headings and links can make a longer transcript easier to use.
For AI search, Google's AI features documentation supplies the relevant principle: important content is available as text. Publishing the video's substantive information in an HTML transcript applies that principle to video. The same documentation also says structured data matches the visible text on the page.
Related terms
FAQ
Can I create a transcript from existing captions?
Yes. W3C's transcript guidance describes using caption files as a starting point. You then add important visual information and format the text for reading.
Does a transcript replace captions for accessibility?
WCAG treats them as distinct provisions. W3C's caption guidance places captions for prerecorded video with meaningful audio at Level A, while its media alternative explanation places Success Criterion 1.2.8 at Level AAA.
Can I use a summary instead of a full text alternative?
A summary alone leaves out information required by W3C's media alternative explanation. It calls for dialogue, relevant sounds and visual descriptions in sequence, with equivalent functionality for interactions.
Sources
- W3C WAI, "Transcripts", updated 17 September 2024
- W3C WAI, "Captions/Subtitles", updated 17 September 2024
- W3C WAI, "Understanding Success Criterion 1.2.8: Media Alternative (Prerecorded)", updated 9 March 2026
- W3C WAI, "Transcribing Audio to Text", updated 24 November 2020
- Google Search Central, "AI features and your website", updated 10 December 2025
Check how a page reads to answer engines. Open the GEO score analyzer