How to read a script on camera without looking like you’re reading
Nobody watching you is hunting for your script. They are watching your eyes, and your eyes will hand it over. Everything below is about that, and software is only one part of it.
To read without looking down, put the words just below the lens, keep each line short enough to read in one glance, and practise the passage aloud before you record. The setup helps keep your eyes near the camera; a relaxed delivery still takes practice.
The glance gives you away
A viewer registers where you are looking faster than they parse what you are saying. Each glance down reads as a half-second in which you left the room, and four or five in a minute make a take feel recited.
Reading has a second signature: talking is a fairly steady gaze, reading is a sweep and a snap back, in a rhythm the eye catches within seconds. That is why the same words look worse beside the lens than above or below it. Vertical movement passes for thinking; horizontal movement reads as reading.
Write something your mouth will agree to
Most of the “I look like I’m reading” problem is made at the writing desk: a page-shaped sentence cannot be spoken at pace without your eyes going back to check it.
- One idea per sentence. If a comma is holding it together, cut it in two.
- Contractions everywhere. You say “it’s”; the draft says “it is”.
- Nothing you must hold in your head to land an ending.
- Break lines where you would breathe, not where the margin runs out.
Then read it aloud, standing, at full volume. Whatever you stumble over twice is a writing problem, and rewriting beats rehearsing around it.
Make the angle smaller
What the camera sees is the angle between the lens and your eyeline, and an angle is cheap to cheat: a ten-centimetre offset is glaring at arm’s length and nearly invisible with the camera three metres back and the shot framed tighter. Put the reading surface directly under the lens, never off to one side, and make the text big enough that a whole phrase lands in one look.
Beam splitters and when they are worth it
A beam splitter is half-silvered glass at forty-five degrees in front of the lens. The camera shoots through it while you read the script reflected off its face, so the words sit on the optical axis and the offset is gone.
The cost is a rig, a tripod, setup time, light lost in the glass, and a script flipped horizontally to read correctly in the reflection. Worth it at the same desk every week, not worth it handheld or once. Longer scripts and rigs are their own subject.
Learn two sentences by heart
The first and the last; read everything between them. The opening decides whether a viewer is being spoken to or read at, and the close catches readers out the other way: there is no next line to reach for, so the eyes drop while the mouth is still finishing.
Pace gives you away too
A conventional prompter runs at a speed you chose before you knew how the take would go, so the performance turns into a negotiation with a machine: hurrying when a line climbs away, padding when it hasn’t come round, never risking the silence before the word that matters. People hear that before they see anything.
Set up a short practice take
Start with a passage of three or four sentences, not the whole talk. Put the phone at eye level and place the reading window as close to the lens as your framing allows. On a vertical phone, try the narrow Strip shape just below the front camera; if its two lines are too few, use Tall. Increase the script size until each short phrase is readable in one glance, and break lines where you naturally breathe. The floating window snaps to a corner and cannot be moved freely or tapped once the take starts, so settle its shape, rotation and margins before you record.
For example, a short product explanation might read:
We cut the setup from ten minutes to two.
The team now checks the same three things before every release.
That means fewer surprises on launch day.
Read those lines aloud once without recording. If “checks the same three things” feels stiff, rewrite it in the words you would use in conversation. Then record a practice take: pause between the first and second sentence, and let the silence stay. In Voice mode the script waits through a pause and follows the words you say; if you repeat a line, it can find that line again. Watch the lens, not the moving text, and use the review to adjust line length, window placement and size before trying a longer passage.
Voice following while recording has been tested in Camera, Instagram and TikTok. If another app takes the microphone, the script waits at your last word and resumes when the microphone returns. Wispim is free for six prompter takes, which do not refill, and scripts are unlimited. The separate browser teleprompter is a different tool with browser speech-recognition limits.
One way to take the chase out
Two paragraphs on the app this site is about. Wispim’s Voice mode recognises what you actually said and scrolls to those words, so pausing, speeding up and repeating a line all behave. There is no 3-2-1 to race. A start delay, ten seconds on a new script, holds the text still before the take starts, in any mode, and after it Voice and Sound wait until you speak. How the following works also covers the less flattering part. Recognition runs on your iPhone where your language supports it. Where it does not, iOS sends the audio to Apple to convert it to text. The language picker marks those languages before the take; a full-screen notice can appear if on-device recognition repeatedly fails and the app falls back to Apple. The floating window shows no notice.
For the angle, the script can sit in a floating window that stays on top of Camera, Instagram or TikTok while you record. It snaps to a corner the way Picture in Picture does, comes in Wide, Tall and Strip shapes, and can be rotated so it reads the right way up under the lens on a rig. Nothing inside it responds to touch. That is an iOS constraint, and the reason spoken commands such as pause and continue exist, in ten languages. Mirror flips the text for the beam splitter above. Wispim runs on an iPhone with iOS 16 or later, on an iPad with iPadOS 16 or later and on a Mac with macOS 13 or later. On a Mac, the floating panel snaps just under the camera, so your eyes stay close to the lens. For a first test, read a short passage in Voice mode and adjust the window before starting.
Questions
How do I keep eye contact on camera if I can’t memorise the script?
Make the angle smaller instead: reading surface directly under the lens, camera stepped back, shot framed tighter. Memorise only the first sentence and the last.
Can people tell when someone is using a teleprompter?
They can tell when someone is fighting one: a sideways sweep at a screen beside the camera, or a delivery locked to a scroll speed. The survey of iPhone teleprompters covers which ones follow a voice instead.
Should I look at the lens the whole time?
No; an unbroken stare reads as odd in its own way. What gives a take away is the same glance, to the same spot, on the same beat.
What if the phone is on a tripod and nobody can press start?
Use a start delay. Wispim offers off, five, ten or twenty seconds before the script moves, in any of the three modes, and shows it inside the floating window as well as full screen.