🎬 Media · Updated October 8, 2026 · 8 min read

Video Captions for Accessibility: A WCAG Checklist

To add captions to a video for accessibility, you need text that is accurate, synchronized with the speech, complete from start to finish, and placed where it does not cover important content. Auto-generated captions are a fast first draft, not the finished product. This guide is for teachers, HR teams, nonprofits and public-sector communicators who publish training, announcements or lectures and need them to work for deaf and hard-of-hearing viewers, as well as for anyone watching on mute. It covers the standard most organizations are measured against, what good captions contain, and a practical workflow with free tools.

🔤 Try AI Studio now — freeOpen →

Lab safety module 3 transcript covering goggles, the eyewash station and common lab mistakes, with SRT and VTT caption exports — The finished transcript with every export option, made entirely in the browser.

💡 Why captions matter more than you think

Roughly one in seven American adults reports some trouble hearing, and captions are the only way many of them can follow a recording at all. They also help people watching in a second language, viewers in a quiet office or a noisy train, and learners who absorb information better when they can read along. On the compliance side, WCAG 2.1 Success Criterion 1.2.2 requires captions for all prerecorded audio in synchronized media at Level A, the most basic tier. Schools, public agencies and many companies use WCAG as their benchmark, and the European Accessibility Act has applied to many consumer services since June 2025.

What WCAG actually asks of your captions

WCAG does not prescribe a font or a color. It asks that the information in the audio is available as text, in time with the picture. Captioning guidelines from the Described and Captioned Media Program and the FCC’s caption quality rules for television fill in the practical detail. Together they point to four qualities:

Accurate: the words match what was said, including names, numbers and technical terms
Synchronous: each caption appears as the words are spoken and stays long enough to read
Complete: captions run from the first word to the last, including the closing remarks
Well placed: text never hides a speaker’s face, a name lower-third or on-screen data

Live streams fall under a separate criterion, 1.2.4, at Level AA. For recorded content, the goal is simple: someone who cannot hear the audio should get the same information as someone who can.

Include more than the spoken words

Captions are not the same as a transcript of dialogue. Deaf and hard-of-hearing viewers also need the sounds that carry meaning. Add them in square brackets on their own line, and keep them short:

Speaker identification when the speaker is off screen or changes, for example [Dr. Patel]
Meaningful sounds such as [phone rings], [applause] or [alarm beeping]
Music cues that set tone, such as [upbeat music] or [music fades]
Tone that changes meaning, for example [sarcastically] or [whispering]

Do not caption every cough or chair creak. Include a sound when a hearing viewer would react to it or when the story depends on it.

Make captions easy to read

Readable captions follow a few well-established conventions. Keep each caption to one or two lines of about 32 to 42 characters, and break lines at natural phrase boundaries rather than mid-phrase. Leave each caption on screen long enough to read comfortably, generally at least a second and a half, and avoid pushing reading speed much above about 160 to 180 words per minute for general audiences.

Contrast matters as much as timing. White or yellow text with a dark outline or a semi-transparent box stays legible over bright and busy backgrounds. If you are styling burned-in captions, check the pairing with the Color Contrast Checker; WCAG asks for a contrast ratio of at least 4.5 to 1 for normal text.

Use sentence case and standard punctuation. All-caps captions are harder to read in long runs, and missing punctuation makes it hard to tell where one thought ends and the next begins.

Draft, correct and deliver with free tools

AI Studio turns the audio into a timed draft on your device and exports SRT or VTT. Read the whole file against the recording, fix names and jargon, add speaker labels and sound cues, then choose how the captions reach the viewer:

Closed captions: upload the SRT or VTT alongside the video on YouTube, Vimeo, LinkedIn or your LMS so viewers can switch them on and resize them
Embedded track: the Video Trimmer can embed a caption track in the file for players that support it
Open captions: burn the text into the picture when the platform has no caption support, such as some social feeds
Transcript: publish a text version as well, which helps screen-reader users and people who prefer to skim

Closed captions are usually the better accessibility choice because viewers can adjust size and color. Burned-in captions reach everyone but cannot be turned off or restyled, so use them where closed captions are not an option.

Step-by-step

1Transcribe the video in AI Studio and export an SRT or VTT file as your first draft.

Draft captions text: 'Welcome to Module 3 of the Lab Safety Course. Before you start any experiment, put on your safety goggles' — Auto-generated text is a first draft — watch the video and correct it before publishing.

2Watch the full video with the caption file open, correcting names, numbers and any misheard words.

The .srt and .vtt caption downloads highlighted for the lab safety module transcript — Download SRT for most players or VTT for the web, then add speaker labels and sound cues.

3Add speaker labels and bracketed sound cues, and split long lines into one or two lines of about 32 to 42 characters.

End of the lab safety transcript: 'a short quiz with 10 questions… when in doubt, ask before you pour' with 'must take three' to correct — Watch the video with the text open and correct lines like 'must take three' (mistake three).

4Upload the file as closed captions where the platform allows it, or burn open captions with the Video Trimmer, and publish a transcript too.

Common mistakes to avoid

⚠️Publishing uncorrected automatic captions, where a misheard drug name or dollar figure can change the meaning entirely.

⚠️Captioning only the dialogue and leaving out a phone ringing or an off-screen speaker that the story depends on.

⚠️Placing burned-in text over a speaker's name bar, a chart or a sign-language interpreter.

⚠️Cramming three long lines into each caption so viewers must choose between reading the text and watching the picture.

Pro tips

✓Build a glossary of names and technical terms for your organization and check every caption file against it.

✓Ask a deaf or hard-of-hearing colleague to review one video; one real viewer finds problems no checklist does.

✓Keep the corrected SRT file in your archive so the next edit or translation starts from clean text.

✓For lectures, announce slide numbers in speech so both captions and transcript can reference them.

✓Test captions on a phone as well as a monitor; lines that fit on a laptop can wrap badly on a small screen.

Frequently asked questions

Are auto-generated captions enough for WCAG?

Usually not on their own. Automatic captions are a good first draft, but WCAG expects captions that convey the content accurately, so a human review for names, numbers and sound cues is needed before you can rely on them.

What is the difference between captions and subtitles?

Captions assume the viewer cannot hear the audio and include speaker labels and sounds. Subtitles usually assume the viewer can hear but needs a translation, so they cover dialogue only.

Should I use closed or open captions?

Closed captions are generally better for accessibility because viewers can turn them on and adjust size and color. Use burned-in open captions when the platform does not support caption files.

Do I need captions for a video with no speech?

If the audio carries information, such as narration-free instructions with meaningful sounds, describe those sounds. If the video has only background music, a short note such as [instrumental music] is usually sufficient.

Is my video uploaded when I create captions?

Not when you transcribe in AI Studio or generate subtitles in the Video Trimmer; both run in your browser. The file only leaves your device when you upload the finished video and captions to your chosen platform.

📌 Bottom line

Accessible captions are accurate, synchronized, complete and well placed: start with an automatic draft, correct every line, add speakers and meaningful sounds, keep lines short and high-contrast, and deliver closed captions wherever the platform allows.

Open AI Studio →