Accessible video is not a niche concern. Hundreds of millions of people are deaf or hard of hearing, millions more watch with sound off or in a second language, and accessibility law increasingly treats online video as covered content. The good news: most of video accessibility comes down to captions done right, plus a few contrast and structure basics. This checklist gives you the practical standard. OneDrop Video Captions (a video captioning tool, unrelated to the One Drop diabetes-health company) is a free browser-based AI video caption generator that handles the captioning half with private transcription and SRT/VTT export.
One important note: this article summarizes widely published accessibility guidance in plain language. It is not legal advice, and it does not certify WCAG conformance for any specific video.
The short answer
An accessible video needs accurate synchronized captions (not raw auto-captions), speaker identification, readable high-contrast caption styling, and a transcript or subtitle file where the platform supports it. Under WCAG, prerecorded video needs captions (1.2.2) and audio description or a text alternative for visual information (1.2.3/1.2.5). Generate captions with OneDrop Video Captions in your browser, review every word, and export SRT or VTT free.
Why do captions matter so much?
Captions are the single highest-impact accessibility feature in video:
Deaf and hard-of-hearing viewers rely on captions for dialogue, and on caption notes like [music] or [applause] for meaningful sound.
Sound-off viewers include commuters, parents, office workers, and anyone scrolling in public. This is often the majority of an audience.
Second-language viewers follow along better with text support, and many prefer it even when they understand the audio.
Search and comprehension improve because captioned content is skimmable and its transcript is indexable text.
Auto-captions alone do not clear the bar. The standard expects accuracy, which means a human review pass over any AI transcript. That review is the difference between "has captions" and "is accessible."
Are your captions accurate and synchronized?
This is the core of the checklist:
Accuracy target. Aim for 99 percent or better on the words that matter: names, terms, numbers, and instructions. Clear speech with a review pass gets there; raw unreviewed auto-captions do not.
Synchronization. Captions must appear as the words are spoken. Lagging or leading captions force deaf viewers to choose between watching the action and reading the text.
Complete coverage. Caption all meaningful speech, including intros, outros, and background dialogue that matters. Do not skip the "boring" parts.
Sound cues. Include non-speech audio that carries meaning: [door slams], [upbeat music], [crowd cheering]. Skip purely decorative audio.
No auto-only reliance. Platform auto-captions are a starting draft, not a finished accessibility feature. Always review and correct.
In OneDrop, every word of the AI transcript is clickable for correction, timing is adjustable, and low-confidence words are flagged for your review. Our guide to generating SRT subtitles from any video free shows the cleanup workflow.
Are speakers identified?
When more than one person speaks, viewers need to know who said what:
Name speakers on their first captioned line and whenever the speaker changes: MARIA: The budget passed. or [Maria] The budget passed.
Off-screen speakers get labeled too: [NARRATOR], [INTERVIEWER].
Consistent formatting matters more than which exact format you pick. Choose one style and keep it through the whole video.
Overlapping speech should be split so each speaker's words are readable, not mashed into one line.
VTT's <v Name> voice spans exist precisely for this; see our SRT vs VTT comparison for the format details.
Do your captions meet contrast and readability basics?
WCAG's text contrast guideline (1.4.3, 4.5:1 for normal text) was written for web pages, but the principle applies directly to captions:
High contrast always. White text with a dark outline or a solid dark background bar reads over any footage. This is the safest default.
Large enough to read. Caption text should be legible on a phone at arm's length. When in doubt, go bigger.
Readable fonts. Simple, bold sans-serif letterforms. Decorative or thin fonts fail real viewers.
No flashing. Avoid rapidly flashing text or backgrounds, which can trigger seizures (WCAG 2.3.1). Animated caption styles are fine; strobing is not.
Positioning. Keep captions where they do not cover important visuals or get covered by player controls. VTT cue positioning helps here.
Should you provide SRT, VTT, or burned-in captions?
For accessibility, toggleable subtitle files (SRT/VTT) are the standard on platforms that support them (YouTube, Vimeo, web players). They let viewers choose caption size and style in their player, which is itself an accessibility feature.
Burned-in captions are better than nothing on platforms that strip subtitle files (TikTok, Reels, Shorts), but they cannot be resized or restyled by the viewer.
Best practice: publish SRT or VTT where the platform accepts files, and use burned-in captions where it does not. OneDrop exports SRT, VTT, and MP4 from the same reviewed transcript, so you can do both without extra work. Our pillar guide on how to caption a video while your file stays on your device covers the private workflow end to end.
What about audio description and transcripts?
Captions cover audio. Viewers who are blind or have low vision need the visual side too:
Audio description narrates important visual information (actions, scene changes, on-screen text) during natural pauses. WCAG expects it for prerecorded video at higher conformance levels (1.2.5).
Full transcripts combine the caption text with descriptions of visual content in one document. They help deaf-blind viewers using braille displays and anyone who prefers reading.
Descriptive scripting is the cheap version: when you narrate your own actions ("I am clicking the blue Export button"), you reduce the need for separate description.
OneDrop's TXT transcript export is a good starting point for building a full transcript: add visual descriptions to the reviewed text and publish it alongside the video.
The downloadable checklist
Copy this into your production notes and run it before every publish:
Captions
-
Every spoken word is captioned, including intros and outros
-
Accuracy reviewed by a human; names, terms, and numbers correct
-
Captions synchronized to the speech, no noticeable lag
-
Meaningful non-speech sounds noted: [music], [applause], [door slams]
-
Speakers identified by name on changes
-
No reliance on raw unreviewed auto-captions Readability
-
High-contrast styling (white text, dark outline or background)
-
Large, bold, simple font
-
Captions positioned clear of key visuals and player controls
-
No flashing or strobing text effects
-
Line length reasonable; broken at phrase boundaries Formats
-
SRT or VTT file uploaded where the platform accepts subtitles
-
Burned-in captions used where subtitle files are stripped (social feeds)
-
Full transcript with visual descriptions published for important content Review
-
Watched once through with sound off, captions only
-
Checked on a phone as well as a desktop
-
Second pair of eyes on anything legal, medical, or instructional
FAQ
What are WCAG requirements for video captions?
WCAG 2.1 Success Criterion 1.2.2 requires captions for prerecorded audio content, and 1.2.4 extends this to live video. Captions should be accurate, synchronized, and identify speakers. This article summarizes the guidance; it is not legal advice.
Are auto-captions enough for accessibility?
No. Raw auto-captions typically fall short on accuracy, especially with names, accents, and technical terms. Accessibility expects reviewed, corrected captions.
Should I use SRT or VTT for accessible video?
Both work. VTT offers positioning and speaker voice spans that help accessibility, and it is the native standard for web players. SRT has wider editor support. Many publishers provide both.
Do social media videos need captions?
Platforms strip subtitle files, so burned-in captions are the practical answer there. They do not offer viewer restyling, but captioned is far more accessible than uncaptioned.
How do I make captions for free?
Use OneDrop Video Captions, a free browser-based AI video caption generator. Transcribe in your browser with no upload, review and correct every word, and export SRT, VTT, or MP4 free with no signup and no watermark.
What contrast do captions need?
Follow the spirit of WCAG 1.4.3: strong contrast between text and background. White text with a dark outline or background bar over any footage is the reliable default.
Accessibility is a workflow, not a feature
Run the checklist on every video and it becomes habit: transcribe, review every word, identify speakers, style for contrast, export the right format for each platform. The private part is easy too. Caption your video accessibly with OneDrop Video Captions: no upload, no signup, no watermark, and subtitle files that meet the standard.