Verify live captions
Description
This rule does not automatically validate anything. It simply reminds you to manually verify that live video broadcasts (provided via <video> elements) include real-time synchronized captions. Success Criterion 1.2.4 Captions (Live) of WCAG requires that all live audio content in synchronized media has captions, except when the media is a media alternative for text and is clearly labeled as such. Live captions must be accurate, synchronized, complete, and available in real time to ensure users who cannot hear the audio have equal access to time-sensitive information such as news streams, conference webcasts, and live events.
Disabilities impacted
- Hearing – People who are deaf or hard of hearing rely on captions as the primary means of accessing audio content. Without real-time captions, live video is effectively silent and inaccessible.
- Cognitive disabilities – Users with cognitive or learning disabilities often benefit from reading captions alongside audio to improve comprehension and retention of spoken information during live broadcasts.
- Situational limitations – Users in noisy environments (e.g., public transit, open offices) or quiet settings (e.g., libraries, late-night viewing) depend on captions when audio cannot be played aloud.
- Auditory processing disorders – Individuals with auditory processing disorders may struggle to distinguish speech from background noise in live audio; captions provide a reliable text-based alternative.
Why it matters
Live video is increasingly used for critical communications: emergency alerts, government press conferences, remote education, corporate town halls, and healthcare announcements. Without real-time captions, deaf and hard-of-hearing viewers are excluded from participating in these events as they happen. Providing live captions is not only a legal requirement under ADA, EAA, and Section 508 but also a fundamental aspect of inclusive design that ensures equal access to time-sensitive information. WCAG also recognizes this through SC 1.2.4, which mandates captions for all live synchronized media. Failure to provide live captions can lead to legal liability, reputational harm, and exclusion of a significant portion of the audience.
Coding problems and solutions
Common coding problems
The following is a checklist of what to manually verify when auditing live video content:
- No caption track element – The
<video>element does not contain a<track kind="captions">child, or the element is absent entirely. This is the most common omission. - Incorrect track
kindvalue – The track useskind="subtitles"orkind="metadata"instead ofkind="captions", which assistive technology may not treat as an accessibility caption source. - Broken or missing track source – The
srcattribute is empty, returns a 404, or points to a static WebVTT file that is never updated during the live event, causing captions to fall out of sync. - Missing language identification – The
srclangattribute is omitted or incorrect, preventing user agents from matching caption language to viewer preferences. - Reliance on auto-generated captions without review – Automated speech recognition may produce inaccurate or delayed captions that do not meet WCAG accuracy requirements without human correction or a live captioner.
How to fix it
Add a caption track to the video element
The simplest and most reliable method is to include a <track> element with kind="captions" as a child of the <video> element. The src attribute must point to a live-capable WebVTT endpoint that streams updated captions during the broadcast.
<video controls width="640">
<source src="live-event.m3u8" type="application/vnd.apple.mpegurl">
<track kind="captions" src="https://live-captions.example.com/stream.vtt"
srclang="en" label="English">
</video><video controls width="640">
<source src="live-event.m3u8" type="application/vnd.apple.mpegurl">
</video>Use a live-capable caption source
For live events, the WebVTT endpoint must serve updated captions in real time. The example below polls a live caption API every 5 seconds and refreshes the text track cues. Always store the interval ID and clean up on page unload.
const trackElement = document.querySelector('track[kind="captions"]');
const textTrack = trackElement.track;
const pollInterval = globalThis.setInterval(() => {
globalThis.fetch(trackElement.src)
.then((response) => response.text())
.then((vttData) => {
while (textTrack.cues.length > 0) {
textTrack.removeCue(textTrack.cues[0]);
}
// Parse vttData and add updated cues
})
.catch((err) => console.warn('Caption refresh failed', err));
}, 5000);
globalThis.addEventListener('beforeunload', () => {
globalThis.clearInterval(pollInterval);
});<video controls>
<track kind="captions" src="captions.vtt" srclang="en" label="English">
</video>
<!-- captions.vtt is a static file never updated during the live event -->Identify the caption language
Always provide the srclang and label attributes on the <track> element so browsers and assistive technology can correctly present the caption language to users. When offering multiple languages, include a separate <track> for each.
<video controls>
<track kind="captions" src="captions-es.vtt" srclang="es" label="Spanish">
<track kind="captions" src="captions-en.vtt" srclang="en" label="English">
</video><video controls>
<track kind="captions" src="captions.vtt">
</video>
<!-- srclang and label are missing; users cannot identify the language -->Provide a transcript as fallback
When captions cannot be embedded in the video stream, offer a link to a live-updating transcript adjacent to the video player. The transcript must also update in real time to be useful.
<video controls>
<source src="live-event.m3u8" type="application/vnd.apple.mpegurl">
<track kind="captions" src="captions-en.vtt" srclang="en" label="English">
</video>
<a href="https://example.com/live-transcript"
target="_blank" rel="noopener">
View live transcript (opens in new tab)
</a><video controls>
<source src="live-event.m3u8" type="application/vnd.apple.mpegurl">
</video>
<!-- No captions, no transcript, no fallback for deaf users -->Known Limitations
- Manual verification required – This rule does not automatically validate caption accuracy, synchronization, or completeness. It only flags
<video>elements for human review. A developer or tester must inspect each flagged element to confirm captions are present and correct. Because the rule is manual-only, it cannot catch issues automatically and depends entirely on developer or tester judgment. - Dynamically added video –
<video>elements injected via JavaScript after the initial page load (e.g., single-page application routers) may not be detected if they fall outside the rule’s initial selector scan. A manual review of dynamic content is essential. - Streaming protocol complexities – Captions embedded in HLS, DASH, or other streaming formats may not use the standard
<track>element. This rule cannot inspect the manifest or segment files for embedded caption tracks. - Cross-origin restrictions – A
<track>element pointing to a cross-origin caption file may be blocked by CORS policies. This rule cannot detect CORS failures. - Live caption quality – Even when a caption track exists, the rule cannot assess caption quality, accuracy, or synchronization. Human review is essential to verify that captions are correctly transcribed and displayed in sync with the live audio.
Resources
- WCAG 2.2 Success Criterion 1.2.4 Captions (Live)
- Understanding SC 1.2.4: Captions (Live)
- MDN: The Track element
- W3C WebVTT: The Web Video Text Tracks Format
- W3C WAI: Captions for Audio and Video
Rule
- Audit: accessibility
- Standard: WCAG
- Level: AA
- Success Criteria: 1.2.4
- Rule ID:
captions-live