In one paragraph
Audio elements must have a captions track is the legacy rule label behind this page, not a complete description of the user's need. Prerecorded audio-only content normally needs an equivalent text alternative, commonly a transcript. According to the supplied WebAuditOS specification, this rule currently produces no report findings. Review audio manually and provide a readable, accurate alternative next to the recording.
Why it matters
A listener who cannot hear a recording needs access to its meaningful content through another form. WCAG 1.2.1 addresses prerecorded audio-only content, with a defined exception where the media is a clearly labelled alternative for existing text.
A transcript should capture relevant speech, speaker identification, and meaningful non-speech information. It need not copy every incidental sound, but must preserve what the recording communicates. A short promotional summary is not equivalent to a full interview or instructional recording.
A timed captions file alone is not necessarily an accessible transcript for an ordinary audio player. Native audio playback does not provide the same visual caption area as video. A custom timed-text interface can be useful when accessible, but verify that users can actually read and navigate the equivalent information.
How WebAuditOS evaluates this finding
The supplied specification describes audio-caption as deprecated or excluded in its tag-based axe execution. It says negative cases would require review rather than produce violations, and the report includes violations only. Consequently, this page must not suggest that today's report confirms missing transcripts or grants a clean bill for audio.
Version and execution configuration must be checked before publishing a current product claim. The historical rule's track inspection is not a review of transcript existence, accuracy, discoverability, or a complete media requirement set. No finding is not a pass for prerecorded audio accessibility.
How to fix it
- Inventory audio-only recordings, including podcast embeds and recordings loaded after interaction. Distinguish them from video with sound, live streams, and clearly labelled audio alternatives to existing text.
- Create an equivalent transcript for each relevant recording. Automated transcription can provide a draft, but verify names, technical terms, numbers, speaker changes, and content that would alter meaning if mistranscribed.
- Make the text easy to find at the player. For example:
<audio controls src="/media/interview.mp3"></audio>
<p><a href="/transcripts/interview">Read the interview transcript</a></p>Use real corresponding content and accessible link wording. This fragment does not demonstrate every player keyboard or naming requirement.
- Prefer readable HTML when suitable so users can search, select, enlarge, and navigate the text. Another format needs its own accessibility review and should not be treated as usable simply because it downloads.
- Keep transcripts aligned with edited recordings. If material is removed or corrected, update the alternative and any time references. Identify the recording clearly when several players share a page.
- If using synchronised highlighting or custom text tracks, test keyboard access, focus behaviour, reading order, and whether the full content remains available when synchronisation fails. Do not replace a usable static alternative with a fragile visual-only widget.
- Review player controls, autoplay sound, and third-party embedded content separately. A transcript does not repair inaccessible play buttons or unwanted audio interfering with screen-reader output.
- Record manual review results and recheck after media changes. Rescanning other rules is useful, but the documented absence of this finding cannot validate the transcript task.
What the fix can break
A transcript can expose speech or personal details that were already audible but not previously searchable. Review publication permission and necessary redaction as part of the content workflow, while keeping the published alternative equivalent to the published recording.
Unreviewed automatic transcription can materially misstate instructions. Assign an editorial owner rather than treating file generation as completion.
Do not spend effort adding an invisible captions track solely to satisfy a legacy label. Deliver the alternative people can actually use.
Questions
Does a transcript replace video captions?
Not generally. Audio-only and video-with-audio have different requirements. Review the actual media type and applicable level.
Why does WebAuditOS show no finding for missing audio text?
The supplied configuration does not surface this rule as a violation. That is a coverage limitation, not evidence of accessibility.
Is live audio covered by the same criterion?
No. Live media has separate requirements and conformance-level considerations. Plan its alternatives separately.
Which related checks help?
Review video captions, autoplay audio, and the automation limits guide.
What severity does this get in the report?
None, because the finding does not appear. Severity for an axe finding comes from the impact axe reports for the element that failed - axe declares this rule critical, which we would map to high. But axe 4.13 marks the rule deprecated and does not execute it in a tag-based run like ours, and even when run, it only ever asks for review.