WCAG 1.2.1 — Audio-only and Video-only (Prerecorded)

What this requires

For prerecorded audio-only content (a podcast, an interview, a voice memo), a transcript or equivalent text alternative must be available. For prerecorded video-only content with no audio track (a screen recording, a silent product demo, a GIF that conveys information), either a text alternative or an audio track that describes the visual content must be available. The alternative has to convey the same information — a one-line summary of a 40-minute podcast is not an alternative.

How AI coding tools fail this

When asked to "embed this podcast" or "add a video player", AI assistants generate the player markup and stop. The transcript is treated as out-of-scope content rather than as a required equivalent. The resulting page renders correctly for hearing or sighted users and contains nothing at all for the rest.

The second pattern: silent screen recordings or product demo GIFs embedded as <video autoplay muted loop> with no caption, no description, and no surrounding text explaining what happens on screen. The visual demonstrates the product; the page tells the screen-reader user nothing.

The third: generated <audio> and <video> elements with controls but no <track> element pointing at a transcript file. The player is accessible enough to operate; the content is not.

Failing examples

Audio embed with no transcript:

<audio controls src="/episodes/12-launch-day.mp3" />

Silent product demo with no description:

<video autoPlay muted loop playsInline>
  <source src="/demo-create-project.mp4" />
</video>

Passing examples

Audio with a linked transcript on the page:

<figure>
  <audio controls src="/episodes/12-launch-day.mp3">
    <track
      kind="descriptions"
      src="/episodes/12-launch-day.vtt"
      srcLang="en"
    />
  </audio>
  <figcaption>
    <a href="/episodes/12-launch-day/transcript">Read the full transcript</a>
  </figcaption>
</figure>

Silent product demo with a textual equivalent:

<figure>
  <video autoPlay muted loop playsInline aria-labelledby="demo-desc">
    <source src="/demo-create-project.mp4" />
  </video>
  <figcaption id="demo-desc">
    The user clicks "New project", names it "Acme site", picks a
    template, and the project opens with three starter pages.
  </figcaption>
</figure>

Edge cases

A transcript is not the same as captions. Captions belong on synchronised media and cover 1.2.2. A transcript is a separate document for audio-only or video-only material.
Decorative audio or video — a background ambient loop with no informational content — doesn't need an alternative, but it must not autoplay (see 1.4.2).
Auto-generated transcripts from a speech-to-text service satisfy the criterion only if they are reviewed and corrected. Raw machine transcripts of multi-speaker conversation are usually wrong enough to fail the equivalent-information requirement.
Animated GIFs that convey information (a "how it works" animation) are video-only content and need a text equivalent. A decorative loop does not.
Sign-language video without an accompanying audio or text track is video-only for non-signers and needs an alternative.

How Jeikin handles this

This criterion is manual — no automated tool can verify that a transcript matches the audio. Jeikin tracks it as a guided review item: the dashboard lists every <audio> and <video> element the scanner finds and asks the reviewer to confirm an equivalent alternative is in place. The evidence is a timestamped record that the check happened, not an automated claim of compliance.