Captions and audio description for small business videos

Updated October 8, 2026 · 7 min read

Short answer: Prerecorded videos with speech generally need accurate captions to meet WCAG 2.1 AA (1.2.2). Audio description (1.2.5) applies when important visual information isn't conveyed by the soundtrack. Auto captions usually need human review before you rely on them.

Video is one of the most common places small business sites fall short on accessibility. Owners often assume that uploading to YouTube or Vimeo takes care of it. It usually doesn't, because the platform's automatic captions are a starting point, not a finished result.

This guide explains the terms, what WCAG 2.1 AA asks of prerecorded video, how to fix auto captions, and when audio description comes into play. It is general information, not legal advice. For legal questions, talk to an attorney.

Captions, subtitles and transcripts: what's the difference?

These terms get mixed up, and the differences matter when you are checking against WCAG.

Term	What it contains	Main audience
Captions	Spoken words plus meaningful non-speech sound (music cues, door slams, "[laughter]") and speaker identification	People who are deaf or hard of hearing
Subtitles	Usually just the spoken words, often translated into another language	Viewers who can hear but don't understand the language
Transcript	A full text version of the audio, and ideally the visual action, on a page	Deaf-blind users (via braille displays), people who prefer reading, anyone who wants to search or skim

A subtitle track that only translates dialogue may not meet the captioning requirement on its own, because it can leave out sounds that matter to understanding the video.

Open vs closed captions

Closed captions can be turned on and off by the viewer. Open captions are burned into the video image. Both can satisfy the requirement, but closed captions are more flexible: users can adjust them, and the text is available to assistive technology and search.

What WCAG 2.1 AA requires for video

The relevant success criteria for video with sound are:

1.2.2 Captions (Prerecorded) (Level A): prerecorded video that has audio needs captions, except when the video is a media alternative for text and is clearly labeled as such.
1.2.3 Audio Description or Media Alternative (Prerecorded) (Level A): prerecorded video needs either audio description or a full text alternative.
1.2.5 Audio Description (Prerecorded) (Level AA): prerecorded video needs audio description. This is the one that applies at the AA level.
1.2.1 Audio-only and Video-only (Prerecorded) (Level A): audio-only content needs a transcript; silent video needs a text or audio alternative.

Live captions (1.2.4) apply to live broadcasts. Most small business sites mainly deal with prerecorded content, so this guide focuses there. You can see the full list of requirements in the WCAG 2.1 AA checklist and the WCAG reference.

Fixing automatic captions

Auto-generated captions are a convenient first draft. They often struggle with:

Product names, brand names and people's names
Industry jargon and acronyms
Accents, background noise and overlapping speakers
Punctuation and sentence breaks
Non-speech sounds, which are usually missing entirely

Treating the auto track as final is a common reason videos fail review. A practical workflow:

Generate or upload a caption file. Most platforms let you start from an automatic track or upload your own.
Edit for accuracy. Play the video while reading the captions. Correct wrong words, names and numbers.
Add speaker labels and meaningful sounds where they help understanding, such as "[phone ringing]" or "[upbeat music]".
Check timing and line length. Captions should appear in sync with speech and be short enough to read comfortably.
Publish the corrected file and confirm it displays in the player your site actually uses.
Re-check after edits. If you re-cut the video, the old caption file may no longer match.

A caption file is typically a small text file such as WebVTT (.vtt). If you host the video yourself with the HTML <video> element, you attach it with a <track> element:

<video controls width="640">
  <source src="/videos/shop-tour.mp4" type="video/mp4">
  <track kind="captions" src="/videos/shop-tour.en.vtt"
         srclang="en" label="English" default>
  Your browser does not support the video element.
</video>

Use kind="captions" rather than kind="subtitles" when the track includes non-speech sounds.

Add a transcript

A transcript on the page near the video is useful even when captions exist. It helps people who use braille displays, those on slow connections, and anyone who would rather read. It also gives you a way to satisfy the alternative-for-media path for some criteria. Keep it as real page text, with headings and structure where it helps, so it works with info and relationships requirements.

Background and decorative videos

Many sites use a muted looping video as a hero background. Whether captions are needed depends on whether the video carries meaningful information.

Purely decorative, no audio: typically no captions or description are required, but it should not distract or block content. If it plays automatically for more than five seconds, WCAG 2.2.2 (Pause, Stop, Hide) requires a way to pause or stop it.
Silent but informative: if the video shows something users need to understand, such as a process or a menu, provide a text alternative. See Non-text Content.
Has audio: treat it like any other video and caption it. Avoid audio that starts on its own, since that interferes with screen reader users.

Make sure any video controls are reachable and operable by keyboard. See Keyboard and Name, Role, Value for custom players.

When audio description applies (1.2.5)

Audio description is narration that describes important visual information that isn't covered by the existing soundtrack. It mainly helps people who are blind or have low vision.

Criterion 1.2.5 requires it for prerecorded video at Level AA. However, the W3C notes that it isn't needed when all the visual information is already conveyed through the audio. Consider these examples:

A talking-head interview where the speaker explains everything: the audio already carries the content, so little or no description may be needed.
A product demo where the narrator says "click here" while the screen shows a menu: important visual information is missing, so description is needed.
A slideshow with music only: the content is visual, so it needs description or a text alternative.

Practical ways to meet it

Write the script to include the visuals. Narrate what is on screen as you speak. This often removes the need for a separate track.
Provide a described version of the video as an alternate track or separate file.
Offer a full text alternative (a transcript including visual details). Note that the AA criterion specifically calls for audio description, so a text alternative alone satisfies the Level A criterion 1.2.3 but may not satisfy 1.2.5 on its own.

If you target WCAG 2.1 AA, plan for audio description wherever visuals carry meaning that the sound doesn't. How courts or regulators treat any specific gap varies, and an attorney can advise on your situation. For background, see ADA website lawsuits explained and what to do with a demand letter.

A short video checklist

Every prerecorded video with speech has accurate, synchronized captions.
Auto captions have been reviewed and corrected by a person.
Names, numbers and technical terms are right.
Non-speech sounds that matter are included.
A transcript is available on or near the page.
Visual-only information is described in the audio or through audio description.
Autoplaying motion can be paused, and nothing plays sound automatically.
Player controls work with a keyboard and a screen reader.
You have checked the result in the actual embedded player, not just the hosting platform.

If your site runs on WordPress, see the notes in WordPress website accessibility, since embed plugins and themes can change how players behave.

Check your work

Manual review is the only reliable way to judge caption quality, since automated tools can detect whether a caption track exists but not whether it is accurate. You can run a free scan to find missing tracks and other common issues, then review the videos yourself.

For official background, see the W3C's Understanding WCAG documentation and the Department of Justice's ADA resources.

Frequently asked questions

Do I need captions on every video on my website?

WCAG 2.1 AA expects captions for prerecorded video that contains speech or meaningful sound, with limited exceptions such as media clearly labeled as an alternative to text. Silent decorative videos typically don't need them. Whether a specific situation creates legal risk is a question for an attorney.

Are YouTube's automatic captions good enough?

Usually not on their own. Automatic captions often misspell names and jargon, miss punctuation and skip non-speech sounds. Review and correct them, or upload your own caption file, before treating the video as captioned.

What's the difference between captions and a transcript?

Captions are time-synced text shown with the video. A transcript is a separate text document or page section containing the same information. They serve different needs, and having both is a good practice.

Do I need audio description for a simple talking-head video?

Often not, if everything important is already spoken. Audio description matters when the video shows information, such as on-screen steps, text or actions, that the soundtrack doesn't convey. Review each video and ask what a listener would miss.

Does adding captions make my site ADA compliant?

No. Captions address one part of video accessibility, and a site has many other requirements. No single change guarantees compliance or protection from a claim. An attorney can advise on legal exposure.

Proof of Access provides technical accessibility assessments. We are not a law firm and nothing on this page is legal advice. If you have received a demand letter or a lawsuit, talk to an attorney.