Captions land in a default position that suits an average video and not necessarily yours, and on a phone they compete with both your face and the platform interface. The fix is direct manipulation: drag the caption on the canvas to reposition the whole track, pinch to resize it, and tap it to open its styling. Position applies to the track rather than one cue, which is what you want, because captions that move around between cues read as broken.
Auto captions have to put the text somewhere before they know anything about your shot. The default is a reasonable guess, and on a specific video the guess lands on your chin, or over the thing you are demonstrating, or exactly where the platform stacks its buttons.
Fixing it should be the most obvious interaction in the app: touch the caption, move the caption.
The three places captions go wrong
Over your face. If you frame tight, or you are shorter in frame than the default assumes, the text sits across your mouth. Distracting in a way viewers feel without being able to name.
Under the platform interface. Every vertical platform overlays its own chrome on the lower part of the frame: the caption of the post, the account name, the buttons up the right side. Text placed near the bottom of your export can end up behind all of it. This is the one that stings, because it looks fine in your editor and broken in the feed.
Over the content. In a screen recording or a demo, the bottom of the frame is often where the thing you are pointing at lives.
This is Apple's Speech framework at work. It offers on-device recognition, which is what keeps the recording on the machine that made it.
Before captions earn you watch time they are an accessibility obligation, and the W3C sets out the bar in captions for prerecorded media.
Move it by touching it
In Zella on iPhone, when a caption is on screen a small handle sits under it, and it answers three gestures:
- Drag to move the caption's position on the canvas. It is clamped, so you cannot lose the text off the edge of the frame.
- Pinch to scale the caption up or down.
- Tap to select that cue on the timeline and open the caption styling controls.
The reason a handle exists at all is that captions are drawn into the frame rather than being a floating UI element, so there is nothing to grab unless the app gives you something. An editor where the captions cannot be touched is not making a design choice, it just never added the handle. Zella did not have one until a field report pointed out that captions were the only thing on the canvas you could not touch.
Why position applies to the whole track
Dragging moves the position for the entire caption track, not the single cue you happened to grab.
That is deliberate and it is the behaviour you want. Captions that sit in a different place from one line to the next read as a glitch, and no viewer interprets it as intentional design. Position is a property of the caption layer, in the same way a font is.
If you genuinely need one line somewhere else, the tool for that is a separate title or callout placed on top, which is a different element with its own position.
Where to actually put them
A few placements that survive contact with real platforms:
- Keep clear of the bottom fifth of a vertical frame. That is where platform chrome lives. The middle-low third is the safest home for captions in a 9:16 video.
- Above the mouth line, or well below the chin. Not across it. If your framing is tight, going higher usually beats going lower.
- Test at phone size. Type that looks balanced on a laptop is often too small in a feed. Pinch it larger than feels right on a big screen.
- Pick a side and stay there. Consistent placement across a series is part of what makes a series look like one thing.
Once placement is settled, choosing the caption style is the next decision, and word by word captions covers whether to highlight per word at all.
Do the placement last
One ordering note that saves rework. Position and size are the last thing to settle, because everything before them can move the text: changing the font changes the line height, editing the transcript changes the line count, and recutting the video regenerates the timings entirely.
Lock the cut, generate captions, fix the text, choose the font, then place and size. Doing it in that order means doing it once.
Frequently asked questions
Can I move captions on iPhone without going into a settings menu? Yes. Drag the caption on the canvas directly. The position applies to the caption track.
Can I make captions bigger? Pinch the caption on the canvas to scale it, or use the size control in the caption styling panel.
Does moving one caption move them all? Yes, and that is intentional. Captions that change position between lines read as broken.
How do I keep captions clear of the TikTok or Reels interface? Keep them out of the bottom fifth of a vertical frame. The middle-low third is the reliable place.
Can I put one caption somewhere different from the rest? Not as part of the caption track. Add a separate title or callout for that line instead.
Why can I not tap my captions in some editors? Captions are drawn into the frame, so they need an explicit handle to be touchable. If an editor never added one, there is nothing to grab.
Make your next video with Zella.
Record, edit and ship on Mac and iPhone - local, private, free to start.
RELATED