Skip to content

Guide

How to add subtitles to a video permanently

By Rakshit · Updated

To add subtitles to a video permanently, burn them in: the words are drawn into the picture itself, so every player shows them and no viewer can turn them off. A separate subtitle file is the opposite, since it travels beside the video. KeyMo burns captions into every clip it exports, in one of 12 styles.

Burned-in versus a separate subtitle file

There are two ways to put words on a video. A separate subtitle file, such as SRT or WebVTT, holds the text and the timings, and the player draws it on top when the viewer asks. A burned-in caption, sometimes called a hard subtitle or open caption, is part of the video frames. Nothing is attached, so nothing can fall off.

The W3C describes the same split: most captions are closed, meaning the viewer can show or hide them, while open captions are always displayed and cannot be turned off. WCAG gives a case of the second kind, captions that are images of text embedded in the video.

What changes when subtitles are permanent
Burned-in (hard)Separate file (soft)
Viewer can turn it offNoYes, if the player has a captions button
Looks the same everywhereYes, it is part of the pictureNo, the player decides the look
Fix a typo laterRe-export the videoEdit the text file and upload it again
Several languagesOne language per copy of the videoOne file per language, switchable
Search and accessibility tools can read the wordsNo, they are pixelsYes, it is text
Works if the platform drops caption filesYesNo

Three places it matters: YouTube, Instagram, TikTok

On YouTube, a separate file is a first-class option. Its help page lists accepted formats, from basic SubRip (.srt) and SubViewer (.sbv) to WebVTT (.vtt) and TTML. Two limits are worth knowing. Plain SRT is accepted in its basic version only, with no style markup recognised, and WebVTT is marked as an initial implementation where styling is limited to bold, italic and underline. If you want a specific look on YouTube, burning it in is one way to control it.

The W3C adds the reason: support for positioning and styling in players is inconsistent, and most web video simply uses the player default, usually white characters in a black box.

On Instagram and TikTok the pages we could read do not describe uploading a separate caption file for an ordinary Reel or post, so we do not claim that either way. Meta's own guide for Reels ads calls on-video captions optional but recommended. For a clip you plan to post on all three, a burned-in version is the one that looks identical everywhere.

What KeyMo does

KeyMo burns the captions into the rendered clip. You pick one of 12 styles, or none, and the words appear in time with the speaker, word by word. The export is a vertical 9:16 file at 1080x1920 and 30 fps, H.264 video with AAC audio, with no watermark on any plan. The captions are not an extra layer you can remove later in another app.

Because the burn-in is permanent in the exported file, pick the style before you download, and keep a plain version if you may want one. Choosing No captions gives you a clip without text. Open a clip and tell the chat editor what to change; edit mode accepts clips from 15 seconds to 10 minutes. Source files stay for 7 days and finished clips for 30, so you can still change a style while the clip is stored.

KeyMo cuts clips from long videos. It is not the tool for adding permanent subtitles to a full-length 16:9 video: its output is vertical clips only. For that job you would use a general video editor or an encoder with a subtitle filter.

Is burned-in the right choice for you?

  • Choose burned-in if you post vertical clips to several platforms and want one look everywhere.
  • Choose a separate file if viewers need to switch captions off, read them in several languages, or if the text must be searchable.
  • Choose both on YouTube only if you are happy that the words may appear twice when a viewer turns on the track.
  • If captions are needed for access, remember that burned-in text is open captioning. WCAG lists open captions as a sufficient technique, but the words still have to be complete and accurate.

Where to go next

If your clips will be dubbed or need toggling, read captions versus subtitles and SRT versus VTT. To see how the 12 looks compare, go to the caption styles gallery. The term itself is defined in the open captions glossary entry.

Sources: W3C WAI: Captions/Subtitles (checked 11 October 2026); W3C: Web Content Accessibility Guidelines (WCAG) 2.2 (checked 11 October 2026); W3C: Understanding Success Criterion 1.2.2 Captions (Prerecorded) (checked 11 October 2026); YouTube Help: Supported subtitle and closed caption files (checked 11 October 2026); YouTube Help: Use automatic captioning (checked 11 October 2026); Meta Ads Guide: Instagram Reels video ads (checked 11 October 2026).

FAQ

Can I remove burned-in subtitles later?

Not from the exported file. The words are part of the frames. In KeyMo you can switch the clip to another style or to No captions while it is still stored.

Will burned-in subtitles work on every platform?

Yes, because they are part of the picture and need no caption track. That is why a burned-in clip looks the same on YouTube, Instagram and TikTok.

Which subtitle file formats does YouTube accept?

YouTube lists SubRip (.srt), SubViewer (.sbv), WebVTT (.vtt), TTML and several broadcast formats. It says SRT and SBV are basic versions only, with no style markup recognised.

Do permanent subtitles count as accessible captions?

Open captions can satisfy WCAG 1.2.2 as a sufficient technique, but they must include the dialogue and important sounds accurately. Automatic text alone is not enough without a review.

Can I upload my own SRT to KeyMo?

That is not a documented feature, so do not plan around it. KeyMo generates its own word-by-word captions and burns them in; there is no claim here of importing or exporting subtitle files.

Does burning subtitles into a video change its resolution?

Not by itself. KeyMo exports every clip at 1080x1920 whether captions are on or off, so the output size does not depend on the caption style.