Accessibility & inclusive video

Accessible Social Media Videos: Captions, Placement, Descriptions and Safety

Accessible social media videos have accurate captions, either uploaded as a caption file or burned into the picture and kept clear of the app's buttons and text, plus post text that describes what matters visually, readable hashtags, and no dangerous flashing. Because many feeds start videos muted, captions also decide whether hearing viewers understand the post at all. Platform features vary and change, so check each platform's current help pages.

9 min read · Updated

What makes a social video accessible

A social video reaches people in a very different setting from a video on your website. It starts in a scrolling feed, often muted, inside an app whose buttons, usernames and captions sit on top of the picture. Viewers decide within seconds whether to stay. The usual accessibility layers still apply, but each one has social-specific traps.

  • Deaf and hard-of-hearing viewers need accurate captions that they can actually read on a phone.
  • Blind and low-vision viewers need the post text to carry what the picture shows, and hashtags and emoji that a screen reader can make sense of.
  • Viewers with photosensitive epilepsy and some migraine or vestibular conditions need protection from flashing and intense motion.
  • Viewers with cognitive disabilities benefit from short, clear speech and on-screen text that matches it.

A broader pre-publication list for any video is in the video accessibility checklist.

Three ways captions reach social viewers

Uploaded caption file
You attach an SRT or similar file where the platform allows it. Viewers can usually turn captions on or off, and the text is real text rather than pixels
Automatic captions
The platform generates captions from the audio. Convenient, but quality varies with audio and language, and editing options differ by platform
Burned-in (open) captions
The text is part of the video picture. Always visible, works everywhere the video is reposted, but cannot be turned off, resized or read by a screen reader

Each approach has a place. Uploaded files are closest to how captions work on a website, and some platforms let viewers adjust size or color. Burned-in captions survive downloads, reposts and apps with no caption support, which is why short-form creators often use them. Automatic captions are better than nothing, but review them if the platform lets you edit, because mistakes in names, numbers and calls to action are common. The trade-offs in general are covered in open vs closed captions.

Some teams do both: a clean caption file for platforms that support it and a burned-in version for those that do not. Avoid uploading a caption file to a video that already has burned-in captions, since viewers who turn captions on then see two overlapping sets.

Designing for muted autoplay

Many platforms autoplay feed videos without sound, and plenty of viewers keep them that way in public or at work. A video that depends on its audio for the first few seconds loses those viewers, hearing or not.

  • Make the opening understandable without sound: a caption on the first line of speech, or a short text frame stating the topic.
  • Do not rely on a sound effect, music drop or spoken punchline to deliver the point.
  • Make sure captions start with the first word, not a second later.
  • If the video has no speech, say so in the post text, or use on-screen text to carry the story, so viewers do not wait for words that never come.

Keeping captions clear of the platform interface

Social apps place their own elements over the video: like and share buttons, usernames, post captions, progress bars and sometimes a product tag. In vertical feeds these usually sit along the bottom and down one side. Captions placed in the usual lower band of the frame can disappear underneath them.

  • For burned-in captions, keep the text well inside the frame, often around the lower-middle of a vertical video rather than at the very bottom edge.
  • Avoid the right-hand edge, where many apps stack their buttons.
  • Check each platform's own safe-area guidance where it publishes one, and test by previewing the post on a phone before publishing.
  • For uploaded caption files, the platform decides placement, and you usually cannot move the text. Keep important on-screen graphics out of the bottom of the frame so captions do not cover them.

The same video often needs different placement for a vertical and a horizontal crop, so plan captions after the final crop is chosen.

Making burned-in captions readable

When captions are part of the picture, nobody can enlarge them or change their color, so the design has to work for everyone at once.

  • Use a plain sans-serif font at a size readable on a small phone held at arm's length.
  • Give text a solid or semi-transparent background box, or a strong outline, so it stays readable over bright and busy footage.
  • Keep to one or two lines at a time. Long blocks of text cover the picture and are hard to read in the time available.
  • Leave each caption on screen long enough to read comfortably. Word-by-word animated captions can be fun, but rapid highlighting is hard for some viewers to follow, so keep the animation simple and the pace calm.
  • Keep the wording accurate. Stylized captions that paraphrase or skip words leave deaf viewers with a different message from hearing viewers.

Written descriptions and post text

Screen reader users usually reach a video through its post text, so that text should carry the essentials of what the picture shows.

  • Describe key visual information in the post: what is being demonstrated, any text shown on screen, and anything the speech does not say.
  • Use alt text where the platform supports it for videos or thumbnails; support varies, so the post text is the reliable fallback.
  • For a video with no speech, such as a product turning on a stand, a one- or two-sentence description in the post gives a blind user the same content.
  • Put important links, dates and prices in the post text as real text, not only on screen.
  • Link a full transcript or descriptive transcript from your profile or website for longer videos.

Hashtags, emoji and special characters

Screen readers read text literally, which turns some social habits into noise.

  • Write multi-word hashtags with a capital letter for each word, for example #AccessibleDesign rather than #accessibledesign. Screen readers are more likely to read the words separately, and sighted readers find them easier too.
  • Put hashtags at the end of the post rather than in the middle of sentences.
  • Use emoji sparingly. Screen readers read each one by name, so a row of ten emoji is ten descriptions. Do not use emoji in place of words.
  • Avoid the decorative bold or script letters produced by text generators. They are made of mathematical symbols and may be read out character by character or skipped.
  • Spell out abbreviations that are not widely known.

Flashing and fast motion

Flashing content can trigger seizures in people with photosensitive epilepsy. WCAG's guidance, commonly summarized as no more than three flashes in any one-second period, is a sensible ceiling for social video too; check the success criterion itself for the exact definition, which also covers flashes that are small or low in contrast. Strobe effects, rapid cuts between very bright and dark frames, and intense red flashes are the usual risks.

  • Remove strobe effects and rapid bright flashes where they are not essential.
  • If flashing is part of the content, such as concert footage, add a warning at the very start of the video and in the post text, before the flashing begins.
  • Be cautious with fast zooms, shaking camera effects and large moving patterns, which can cause discomfort for people with vestibular disorders.
Hypothetical example: a museum's event reel

A museum posts a 40-second vertical reel of a late opening. The first version opens on music with no speech, has captions at the very bottom that the app's username covers, includes a strobe-lit dance floor and ends with the hashtag #museumlatesnight. The revised version opens with on-screen text naming the event and date, moves burned-in captions up into the lower-middle of the frame with a background box, adds a flashing-lights warning to the first second and the post, describes the exhibits shown in the post text and uses #MuseumLatesNight.

Mistakes and limits to keep in mind

  • Trusting automatic captions without checking, especially for names, brands and calls to action.
  • Burning in captions and then also uploading a caption file, so two sets overlap.
  • Captions that cover faces, on-screen text or the app's own controls.
  • Treating captions as a design element and paraphrasing speech for style.
  • Assuming one export works on every platform. Interfaces, caption support and safe areas differ and change, and publishing language versions across platforms has its own considerations, covered in publishing translated videos across platforms.
  • Captions on social media still serve speech. Important sounds, such as an alarm in a safety demonstration, need a caption like [alarm sounding] or a mention in the post.

Using mydubly for social video captions

mydubly can draft the caption text. Upload the finished clip and you get SRT and VTT files plus a transcript, in the spoken language or a target language, with one subtitle track per job. A 40-second clip is charged at the minimum of 5 credits (0.5¢). The SRT can go straight to platforms that accept caption files, or into an editor that turns subtitle files into burned-in captions.

Be clear about what the files are. Cues are single lines, numbered in order, one per recognized segment, with no styling or positioning, and mydubly does not burn captions into the video. The files contain recognized speech only: no sound tags, no speaker labels and nothing about visual content. For accessibility use, review the text, add important sound cues, and check timing on the platform itself; an SRT file can be edited in any text editor. For finding short clips inside a long recording first, see finding clips with transcripts, and the subtitle generator page explains the outputs.

Next step: fix your next three posts

Before your next three video posts, check four things on a phone with the sound off: the opening makes sense without audio, captions are accurate and clear of the app's buttons, the post text describes what matters visually with readable hashtags, and nothing flashes rapidly. The guide to adding subtitles to a video covers attaching caption files and burning them in.

Frequently asked questions

Are burned-in captions or uploaded caption files better for social media?

It depends on the platform and how the video will travel. Uploaded files can be turned off and are real text, while burned-in captions stay visible when a video is downloaded or reposted to apps without caption support. Many teams publish a burned-in version where caption support is weak and a caption file where it is good.

Are automatic captions on social platforms good enough for accessibility?

They are a starting point. Quality depends on the audio, accents and language, and errors in names, numbers and key messages are common. If the platform lets you edit automatic captions, review them; otherwise upload your own reviewed file or burn in captions you have checked.

Where should captions go on a vertical video?

Away from the very bottom and the right-hand edge, where most apps place usernames, post text and buttons. Many creators place burned-in captions in the lower-middle of the frame. Check each platform's safe-area guidance where available and preview the post on a phone before publishing.

Why should hashtags use capital letters?

Capitalizing each word in a multi-word hashtag, such as #AccessibleDesign, helps screen readers read it as separate words instead of trying to pronounce one long string. It also makes the hashtag easier for sighted people to read, including people with dyslexia.

How do I make a video without speech accessible on social media?

Describe what happens in the post text, put any important on-screen information there as real text, and use alt text if the platform supports it. If music or sounds carry meaning, mention them too. A deaf viewer may need captions for important sounds even when nobody speaks.

Do I need a warning for flashing lights in a social video?

If the video contains strobe effects or rapid bright flashes that cannot be removed, add a warning at the start of the video and in the post text before the flashing begins. Removing or reducing the flashing is safer still, since warnings are easy to scroll past in a feed.

Can mydubly add captions directly to my social video?

No. mydubly produces SRT and VTT files and transcripts but does not burn captions into the picture. You can upload the SRT to platforms that accept caption files, or import it into a video editor to create burned-in captions, after reviewing the text and adding any important sound cues.