most teleprompter apps put the live camera feed behind your scrolling text. that is the single worst place to put it. a wall of moving words sits right where your eyes need to land, and your gaze drifts off the lens. this is a small design choice, and it quietly ruins the one thing talking-head video is judged on.
why behind-the-text is the wrong default#
the appeal is obvious. you want to see yourself while you read, so the app fills the frame with your face and floats the words on top. it looks like a real teleprompter in a studio.
but it isn't one. a studio prompter shows the words over a sheet of glass in front of the lens. your eyes read at the glass, and the lens is right there behind it. the eye-line stays true.
a phone running camera-behind-text does the opposite. the words are the layer in front, and they move. you read at the text, and the text is not the lens. worse, the text is constantly scrolling, so your eyes track it line by line. the camera catches all of that motion.
so the most distracting possible element, a scrolling block of words, gets stacked exactly where you most need a calm, fixed point to look at. that's backwards.
what gaze drift looks like on screen#
at arm's length, off-lens gaze is visible. the viewer can't name it, but they feel it. you look like you're reading, or talking to someone just past their shoulder, or checking your own reflection. all three break the one-to-one feeling that makes a piece land.
the fix is geometry, not willpower. you can't stare down a full-screen scroll and win. the words have to be small, and they have to sit as close to the lens as you can get them.
the closer your reading target is to the lens, the smaller the angle between "where you're looking" and "where the camera is." shrink that angle enough and the drift disappears on screen. that is the whole trick. everything else is in service of it.
the fix: a tile you place, or no preview at all#
open prompter doesn't force the camera behind your text. the camera is optional, and where it goes is your call.
turn it on and you get a small picture-in-picture tile. it's draggable, so you put it where it helps instead of where the app decided. drag it up under the front lens, near the top of the frame, and your eyes barely leave the camera. tap the tile to promote it to a full-screen "how do I look?" preview, check your framing and light, then tap minimize to drop it back to a small tile for the take.
the words still scroll, but they're below a small fixed tile near the lens, not stacked behind a full-frame feed. your reading target is tiny and parked next to the camera.
and if you read better without seeing yourself at all, turn the camera off. the chip cycles off, pip, off. no preview, no tile, nothing pulling your eyes around. plenty of people deliver a cleaner take blind, and the app lets them.
where to put the tile#
a few placement notes that make the difference real.
- top of the frame, under the lens. the front lens sits at the top edge of the phone. park the tile right beneath it so your eyes stay high, close to the camera. a tile in the bottom corner just trades one bad eye-line for another.
- keep it small. the point is a fixed reference, not a mirror you study. a small tile is enough to confirm you're in frame and lit. resist the urge to blow it up mid-take.
- frame first, then shrink. tap to expand full-screen while you set headroom and light. get it right, then minimize. don't fuss with framing on a tiny tile.
when to skip the preview entirely#
sometimes the best preview is no preview.
if you tend to perform to yourself, the tile becomes a distraction of its own. you start fixing your hair, watching your mouth, drifting off the lens to look at the lower-corner version of your face. that defeats the purpose.
turn the camera off and lean on voice tracking instead. it listens to your delivery on-device and scrolls to keep your place, so you don't need a thumb on the screen or an eye on a preview. nothing leaves the phone. with the camera off and the words advancing on their own, there's nothing left to stare at but the lens. that's the goal. here's how voice tracking and the reading lines work.
the extreme case: a beam-splitter rig#
if you want the eye-line dead perfect, a beam-splitter rig puts the words on glass directly in front of the lens, the way studio prompters do, and open prompter drives it with mirror mode and a Bluetooth remote. we wrote a full guide to that setup.
but you don't need a rig for clean eye contact. a tile near the lens, or no preview and voice tracking, gets you most of the way there with the phone you already own.
the short version#
eye contact is the whole game, and behind-the-text fights you on it. so don't fight back. put a small tile near the lens, or turn the camera off and trust your read. the words belong near the camera or out of sight, never stacked over your face.
open prompter is free and open source. if it saved you a take that looked like you were reading, that's the point.