WCAG 1.2.8 — Media Alternative (Prerecorded)

What this requires

For prerecorded synchronised media at Level AAA, a media alternative for time-based media must be provided. This is a full text document that conveys all the information in the video — dialogue, on-screen text, visual actions, sounds, music cues — in a form readable by a deafblind user using a refreshable braille display. It's the most demanding of the media criteria because it can't lean on either the auditory or the visual channel.

How AI coding tools fail this

AI tools generally don't produce these documents; the failure is in how they're surfaced. When a media alternative exists, code-generation tools often bury it as a footer link or omit it from the player UI entirely. The user who needs it most can't find it.

The second pattern: presenting an editorial recap or a marketing summary as if it were a media alternative. A "what we covered" blog post is not equivalent content. The media alternative has to reproduce the information, not interpret it.

Failing examples

A video with no linked text alternative at all:

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

A "recap" passed off as a media alternative:

<a href="/blog/launch-keynote-recap">Launch keynote — recap and highlights</a>

Passing examples

A linked, full text alternative on the same page as the video, covering dialogue, visuals, and sound:

<figure>
  <video controls>
    <source src="/launch-keynote.mp4" type="video/mp4" />
    <track
      kind="captions"
      src="/launch-keynote.en.vtt"
      srcLang="en"
      default
    />
  </video>
  <figcaption>
    <a href="/launch-keynote/transcript">
      Full text alternative — dialogue, visuals, and on-screen text
    </a>
  </figcaption>
</figure>

A sample of the alternative content:

00:00 — Title card on a black background: "Jeikin — launch keynote".
00:04 — [upbeat synth music fades in]
00:08 — SARAH walks onto a small stage. She wears a blue sweater. The
        backdrop is a projected screen showing the Jeikin logo.
00:14 — SARAH: "We started Jeikin because every accessibility tool we
        tried treated the developer like a chore to be policed..."

Edge cases

Captions + audio description ≠ media alternative. A deafblind user can't perceive either. The text document is the only path.
The link must be discoverable to a screen-reader user before they hit play. Putting it inside <figcaption> adjacent to the video is reliable; a deep-buried "Resources" page is not.
Auto-generated transcripts are not adequate by themselves — they miss the visual track entirely.
Pages of long-form video (lecture series, conference archives) often satisfy this via shared transcript pages, which is fine as long as the link is right next to each video.

How Jeikin handles this

This is a manual review item. The scanner records each <video> and asks the reviewer whether a media alternative exists and is linked adjacent to the video. The dashboard tracks the answer per video, with a reminder for projects targeting AAA that a captions-only setup is not enough.