You position your captions in an editor, on a clean 1080×1920 frame, and they look right. Then you post, and TikTok puts a username, a caption, an audio strip and a progress bar across the bottom third, plus a column of buttons down the right. Your last line is underneath all of it.
Nobody publishes the numbers for this, and they change between releases. Here are the ones we use, measured off the current apps and expressed as a percentage of frame height or width so they hold at any resolution.
What each app takes
| Top | Bottom | Right rail | |
|---|---|---|---|
| TikTok | ~10% | ~20% | ~21% |
| Instagram Reels | ~11% | ~25% | ~20% |
| YouTube Shorts | ~8% | ~18% | ~19% |
| Snapchat Spotlight | ~11% | ~23% | ~20% |
Two things fall out of that table immediately.
Reels is the binding constraint at the bottom. Its caption-and-controls stack is the deepest of the four, at about a quarter of the frame. Anything safe on Reels is safe on the other three; anything tuned to TikTok's shallower bar will be covered on Reels.
The right rail is worth more than it looks. A fifth of the width, running most of the height. Full-width caption text runs straight into it, which is why the default line width here is 82% rather than 100% — it clears the buttons on all four without you thinking about it.
Design for the worst case on every edge, not the average. Something safe on average is still covered somewhere, and "somewhere" is wherever your best post ends up.
Where to actually put them
The band between roughly 12% and 74% from the top is clear on all four platforms. That's a lot of room, and the useful part of it is lower than people expect.
- Too high and the captions fight the subject's face, which is usually in the upper middle of a talking-head shot.
- Too low and Reels eats them.
- Around 60–70% down is the sweet spot for most vertical video: below the face, above the furniture.
If your caption block is two lines it occupies roughly 11% of the height, so a block whose top sits at 68% ends at 79% — which clears TikTok and Shorts and collides with Reels by a few percent. Move it up six points and it clears everything. That specific arithmetic is what the safe-area tool does for you.
The mistake worth avoiding
Don't reframe a 16:9 video to 9:16 and then place captions. Do it the other way round.
Cropping a landscape video to vertical throws away about two thirds of the picture and enlarges what's left nearly 1.8×. If you position captions first and crop second, they end up wherever the crop leaves them — often half off the edge. Reframe, check the result against the safe areas, then caption.
Both of those are free tools on this site, and neither uploads your video anywhere. That isn't the point of this post, but it would be strange not to mention it.