Recording guide
A teleprompter for YouTube videos
Nobody delivers eight minutes to camera in one clean pass, and you should not try. The useful question for long-form is not how to avoid mistakes — it is how to make a mistake cost you fifteen seconds instead of the whole take.
Long-form breaks differently
A thirty-second piece you simply record again. At eight minutes that stops working: the fluff always seems to land around minute six, and re-running the whole thing to fix it burns your voice and your patience long before it produces a better take. Most people respond by cutting heavily in the edit, which is how a talking-head video ends up with a jump cut every four seconds.
The alternative is to record long-form as what it actually is — a run of sections, joined. Pause at the end of a thought, breathe, and resume on the exact word. Fluff a line and delete just that part; the script rewinds with it and you carry on from there. What comes out is still one video, so the joins are yours to place rather than artefacts you have to hide.
Read at the lens, not beside it
The script sits high on screen under the camera, over the live picture. On long-form especially, an eye-line that drifts off-axis for eight minutes is far more obvious than the same drift for thirty seconds.
Set the pace, then adjust it
Scroll speed and text size change while you record, and the pace coach shows whether you are running ahead of the script or behind it. Over a long take, drift compounds — catching it at minute two is much cheaper than at minute seven.
Takes stay side by side
Record the intro three ways and compare them before committing. The opening is worth more attention than the middle, and it is the part most likely to need several attempts.
Bigger screen, if you have one
Long-form usually gets shot at a desk rather than at arm's length. Recito runs as a resizable window on macOS with scripts and takes shared across devices — see Recito for Mac if that is your setup.
The free tier will not cover this
Worth saying before you start rather than after: scripts are capped at 750 characters on the free tier, which is roughly forty-five seconds of speech. Any real long-form script runs past that in the first minute, and free saves also carry a small watermark. Short-form creators can genuinely live on the free tier — long-form cannot. The unlock is one payment, not a subscription, but you should know it is required here.
Writing a script you can actually read
Speaking to camera runs around a hundred and fifty words a minute, so an eight-minute video is about twelve hundred words. Two things make that readable aloud: short sentences, and paragraph breaks where you intend to breathe. Prose written for the eye has long clauses that work fine on a page and strand you halfway through on camera — read the script out loud once before recording and cut whatever you stumble on. That pass costs five minutes and saves takes.
What it does not do
Recito records; it does not edit, add captions or chapters, mix audio, or upload. It does not follow your voice — you set a pace and read to it. It will not fix a script that is too long, and it does not light you or improve your microphone, both of which matter more to a long-form audience than anything on screen.