Burned-In or Separate Subtitles: What Platforms Need

Every captioned video ships in one of two shapes. Either the words are drawn into the picture and travel as part of the image, or they live in a separate file the player draws on top. Creators call the first hardcoded, hardsub, or burned in, and the second soft subtitles or a subtitle track. Which one you need is decided by where the video is going, not by preference.

The rule that settles most of it

Short-form social feeds take a video file and nothing else. Upload an MP4 and an SRT to Instagram Reels or TikTok and only the MP4 becomes the post, so the captions are simply gone. If you want words on a Reel, they have to be in the picture when it uploads.

YouTube is the opposite case. It accepts a real subtitle track, keeps it toggleable, and lets you add several languages to the same video. Web players built on the HTML video element do the same thing with a WebVTT file.

So the practical answer for most creators is not one or the other. It is both, from one reviewed source: a captioned video for the feeds, a subtitle file for YouTube, the web, clients and the archive.

What you give up by burning them in

• The viewer cannot turn them off. Anyone watching with sound, or reading a different language, is stuck with them.
• You cannot fix a typo without going back to the edit and exporting the video again.
• Search engines and platforms cannot read the words. Text in a subtitle file is machine readable; text painted into a frame is not.
• A second language means a second export rather than a second track.

What you give up with a separate file

• Styling is mostly out of your hands. The player decides the font, size and position, so the deliberate look of a short-form caption is not available.
• Any destination that ignores the file shows no captions at all, silently.
• The file can be separated from the video in a handoff and arrive without it.

Choosing per destination

• Instagram Reels, TikTok, and short vertical feeds: burned in.
• YouTube, including Shorts uploaded through Studio: subtitle track, and a burned-in version too if the caption styling is part of the edit.
• A website or documentation player: WebVTT track.
• A client or broadcaster deliverable: send what the brief asks for, and keep the SRT even when they asked for a burned-in cut.
• Your own archive: always keep the subtitle file. It is the version you can edit, translate and reuse.

Keep the editable source, whatever you ship

The mistake that costs the most time is treating a burned-in export as the master. Once the words are pixels, the only way to change them is to rebuild the video. Keep the timed text as the source, and treat every burned-in video as an output of it.

That is the workflow the AutoCaption caption generator is built around: review the wording and timing once, then export SRT, VTT, plain text, or a styled captioned video from the same reviewed track. If you already have a subtitle file and need it in the other format, the SRT to VTT converter handles that, and the SRT and VTT guide covers which one a given destination wants.