Which one to use depends on where the video is going, and the answer changes from platform to platform. The short version: open for social feeds, closed for YouTube and your own site, often both.
What are open captions?
Open captions are drawn onto the video frames when the video is exported. You'll also see them called burned-in captions, hardcoded subtitles or hard subs. Once exported, they're pixels, not text.
So they appear on every platform and player, including ones with no caption support at all, and they look exactly the way you designed them, down to the font, position and word-by-word highlight. But the viewer can't switch them off, make them bigger or swap the language, and the platform can't treat them as text.
Most captioned short-form video uses open captions. Feeds often autoplay with the sound off, and a caption the viewer has to switch on is a caption most viewers never see.
What are closed captions?
Closed captions are stored as text with timestamps, usually in an SRT or VTT file uploaded next to the video, or as a track inside the video container. The player draws them over the picture when the viewer asks for them.
Because they're text, a viewer can turn them on and off, and many players let them change the size or color to suit their eyes. One video can carry several language tracks, and the platform can reuse the text, for example to show a transcript or offer machine translation. The downside: they only show up where the player supports them, and the player decides how they look. Broadcast TV, streaming services, YouTube, Vimeo and most course platforms all use closed captions.
Side by side
| Open captions | Closed captions | |
|---|---|---|
| Viewer can turn them off | No | Yes |
| Viewer can change size or color | No | Often |
| Works on every platform | Yes | Only where the player supports it |
| Several languages | One per video file | Yes, one track each |
| Platform can use the text (transcript, translation) | No | Yes |
| Styling | Exactly as you designed it | Up to the player and viewer |
| Delivered as | Part of the MP4 | SRT, VTT, SCC and other formats |
Which one should you use?
For social media clips, use open captions. On TikTok, Reels, Shorts, X and LinkedIn, the text has to be there before anyone taps anything.
For YouTube, Vimeo and your own site, use closed captions, and upload a checked SRT rather than relying on the automatic track. Viewers who need captions can size them the way they like.
For presentations and events, go open. You don't know what player or projector the file will end up on, and a missing caption track in a conference hall is hard to fix on the spot.
Often the answer is both. Post the open-caption version to social and upload the clean original with an SRT to YouTube. Both come from the same transcript, so it's one round of editing, not two.
What accessibility guidelines actually say
People often assume accessibility rules require closed captions. For web content, that isn't quite right. The Web Content Accessibility Guidelines (WCAG 2.x) have Success Criterion 1.2.2, Captions (Prerecorded), a Level A requirement that prerecorded video with audio has captions. W3C's own explanation lists both open captions and closed captions as ways to meet it. What matters is that the captions are there, in sync, and complete, which means dialogue plus speaker identification and meaningful sounds like a door slam or a phone ringing.
So an open caption that covers all of that passes. One that only transcribes speech, or drops half the lines to fit a design, doesn't. Closed captions still have a practical edge for many viewers because they can be resized, but that's a usability point, not a pass or fail rule in 1.2.2.
US television is a separate world. In 2014 the FCC adopted caption quality standards for TV programming covering accuracy, synchronicity, completeness and placement. Those rules are about broadcast and cable programs, not a clip you post to Instagram. Whether any law requires captions on your video depends on who publishes it and where, so if you work for a school, a public body or a large company, check its accessibility policy. Most of those point to WCAG.
Open and closed apply to subtitles too
"Open" and "closed" describe how the text is delivered, not what's in it. Translated subtitles on a foreign film can be burned in (open) or selectable (closed), same as captions. The difference between captions and subtitles, and where SDH fits, is its own topic, covered in our guide to closed captions vs subtitles and SDH.
How to make either kind for free
Both come from the same starting point: an accurate, timed transcript. Our auto captions tool writes one from your video, you fix the words, and then you choose the output. Export an MP4 to get open captions drawn into the picture, or download an SRT or VTT to upload as closed captions. If you already have a caption file and only want it on the picture, the burn-in tool skips the transcription step. If you need the file on its own, use the SRT generator.
One practical tip: our tool writes what's said. It doesn't add sound descriptions or speaker names on its own, so if you're captioning for accessibility, type [door slams] or the speaker's name into the lines where they matter before you export.