Burned-in captions or platform captions
Burned-in captions cannot be switched off, and the text in them is a picture rather than a caption track a platform can correct, translate or search; the format decides which you use.
There are two ways to put words on a video and people mostly pick one out of habit. The habit is usually whichever their editing tool made easiest, which is not a reason.
The decision rule, before the reasoning. Short vertical video, watched in a feed by people who did not choose it, much of it on mute: burn the captions in. Long-form video, where accessibility and search matter and the viewer chose to be there: use a proper subtitle file. Publishing the same content in both places: do both, once each, and stop treating them as the same asset.
What is the difference, exactly?
Burned-in captions are painted into the picture. They are part of the image, in the same way your face is, and they cannot be removed by anybody at any point. You control exactly how they look.
Platform captions come from a separate text file, usually an SRT, which is nothing more than a list of lines with a start and end time attached. You upload it alongside the video and the platform draws the text itself. The viewer can turn it off, the platform can translate it, and a search engine can read it.
Same words on screen. Completely different objects.
What do you lose when you burn them in?
Four things, and none of them shows up on the day you publish.
The viewer cannot turn them off, which matters to people watching with sound who find text distracting, and it matters when your captions turn out to have an error in them.
The burned text itself carries no translation. A platform can still transcribe your spoken audio and offer automatic subtitles from that, but the words you painted on are part of the picture, so the clean, correctable track a viewer switches into another language is the one you did not provide, and that audience is invisible to you because they leave without comment.
They give the platform no clean transcript. The words are pixels, so instead of a corrected caption file it can read, the platform is left working from its own automatic transcription of your audio, mistakes in your vocabulary and all.
And they are permanent. Wrong name, wrong spelling, wrong term, and the fix is a re-export and a re-upload, which for a video with any history attached is a real cost.
What do you lose when you do not?
Reach in a feed, mostly, and a bit of control.
Platform captions look like every other video on the platform, because they are drawn by the platform. On a feed video that is a genuine disadvantage: you have given away the one part of your frame that could have carried your typeface and your styling, and the caption will sit wherever the platform decides, which may be over the thing you are demonstrating.
They also depend on the viewer switching them on, and on the platform's own automatic transcription being accurate, which for anybody with an accent, a specialist vocabulary or a quiet recording it frequently is not. Upload your own file rather than relying on the automatic one, and correct the words that matter to your subject.
Do subtitles help you get found?
A subtitle file gives a platform a clean, timed transcript of everything you said. Burned-in captions give it a guess.
That is a real difference and it is worth being unexcited about. It will not rescue a video nobody wants to watch, and it is not a trick. It simply means the platform knows what your video contains, which is more than it knew before, and for anybody teaching a specialist subject where the vocabulary is the whole point, it is a straightforward thing to stop leaving on the table.
When is the right answer none at all?
More often than you would think, on long-form video, and this means no burned-in set rather than no captions.
If the platform's own captions are accurate on your content, adding a burned-in set gains you very little and costs you everything in the list above. Sometimes the honest answer really is that the native captions outperform anything you would burn in, and the right move is to leave the picture clean and spend the time correcting the transcript instead, so an accurate track is still there for the people who need it.
Check before you decide. Watch one of your videos with the platform's automatic captions turned on and read them. If your subject vocabulary survives, you have less work to do than you assumed. If your subject vocabulary comes out as something unrelated, that is your answer and it is a file you need to upload rather than text you need to paint on.
So which one, in the end?
Short vertical video: burn them in, every time. It is watched in a feed by people who did not choose it, a large share of them on mute, and the styling is part of how anybody recognises your work. Even where the sound is on, the burned-in line is what holds the eye in the first second. Design it properly and keep it clear of the platform's own buttons.
One caveat that outranks the styling: burned-in text is invisible to a screen reader and to anyone using the platform's own caption controls, so where the platform also lets you attach a caption track, add one alongside the burned-in version. Captions are an access requirement before they are a reach lever, and the cheapest way to meet it is to export the caption file from your editor, or upload the platform's automatic transcript and correct the words that matter, so an accurate text file exists.
Long-form video: a subtitle file, every time. The viewer chose to be there, they can switch it on, the words are searchable, and somebody watching in another language is not excluded by a decision you made in an export dialogue.
The one thing not to do is the compromise, which is burning captions into your long-form video because the short-form workflow already did it. That is not a decision, it is a habit that leaked between two formats.
What if you publish the same content twice?
Then it is two assets and it always was. The vertical cut gets burned-in captions built for a muted feed. The long version gets a subtitle file and a clean frame. That is one extra export and one file upload, and it is the point at which most people notice that they have been treating a publishing decision as an editing preference.
Keep reading
This is a service, and the method is written down.
Everything above came out of doing the work rather than writing about it. If you want the method instead of the story, it runs in order on one page.
Start with a free audit
Tell us where your content is now. We will come back with what we would change and what result to expect.
A person reads the channel and writes the audit by hand: a considered read typically takes three working days. That is the usual shape, not a promised turnaround. We use these details only to reply to you: no lists, no lurking.
What you will get
A fit snapshot: where your channel stands, and whether we are a match.
Two to three opportunities: specific, prioritised, yours to keep.
A recommended next step, even if that step is not us.
The audit is free and commits you to nothing: nobody follows up with a call you did not ask for.