WCAG 1.2.3 — Audio Description or Media Alternative (Prerecorded)

What this requires

For prerecorded synchronised media at Level A, the author has a choice: provide an audio description that narrates the on-screen visuals during the natural pauses in dialogue, or provide a full text alternative that covers everything happening on screen — visual actions, on-screen text, scene changes, gestures, anything a blind user would otherwise miss. The text alternative is a complete equivalent, not a summary.

How AI coding tools fail this

When asked to embed a video tutorial or product demo, AI tools generate the player markup and never raise the question of an audio-described version. Even when the visuals carry most of the information (a screen recording of a UI walkthrough), the assistant ships a single audio track and considers the work done.

The second pattern: a <video> with captions but no description track. Captions cover 1.2.2 and the assistant treats accessibility as a single box that has been ticked. Blind users get the dialogue and lose everything happening on screen.

The third: text alternatives that are marketing summaries ("Watch our launch story") presented as if they were equivalents. The criterion requires the same information, not the same intent.

Failing examples

Video with captions but no description and no text alternative:

<video controls>
  <source src="/product-demo.mp4" type="video/mp4" />
  <track kind="captions" src="/product-demo.en.vtt" srcLang="en" />
</video>

A "summary" passed off as a text alternative:

<a href="/blog/launch-recap">Read our launch recap</a>

Passing examples

Either provide an audio-described version of the video:

<video controls>
  <source src="/product-demo.described.mp4" type="video/mp4" />
  <track
    kind="captions"
    src="/product-demo.en.vtt"
    srcLang="en"
    default
  />
</video>

Or provide a full transcript that covers the visuals and the audio:

<figure>
  <video controls>
    <source src="/product-demo.mp4" type="video/mp4" />
    <track kind="captions" src="/product-demo.en.vtt" srcLang="en" />
  </video>
  <figcaption>
    <a href="/product-demo/transcript">
      Read the full transcript with visual descriptions
    </a>
  </figcaption>
</figure>

The transcript reads:

[00:02] The dashboard loads. A sidebar on the left lists three projects.
[00:04] SARAH (voiceover): Start by opening the project you want to review.
[00:06] The cursor hovers over "acme-site" and clicks. The project page opens...

Edge cases

Talking-head videos where the speaker is centre-frame and nothing else is happening on screen often satisfy this criterion via the captions alone — but only if there's genuinely no visual information being conveyed.
Screen recordings and tutorials almost always need either audio description or a transcript. The information is in the screen.
1.2.3 is Level A; 1.2.5 is Level AA. 1.2.3 accepts a text alternative or audio description. 1.2.5 requires audio description (no text-alternative fallback).
Extended audio description (pausing the video to fit longer narration) is a Level AAA criterion (1.2.7).
Embedded YouTube/Vimeo videos inherit the player's accessibility, but the responsibility for the description track or transcript stays with the publisher.

How Jeikin handles this

This is a manual review. The scanner lists every <video> element on the site and asks the reviewer to record which alternative is in place — described version, transcript, or "visuals carry no information". The dashboard keeps the answer as evidence, not as an automated claim.