You record a two-minute walkthrough of your app at full resolution, watch it back on the same screen you recorded it on, and it looks fine. Then you post it. In a feed it plays at a third of the width, on a phone it plays at a sixth, and the button you spent forty seconds explaining is now eleven pixels wide. Nobody can see what you are pointing at, so nobody watches to the end.
The fix everyone lands on is the same: zoom in on the part that matters. The question is how, because the obvious way is slow enough that most people do it once and never again.
Why a full-screen recording reads as noise
A 1440p capture carries roughly four million pixels. The player in a social feed gives it about 500 pixels of width, and a phone gives it less. Everything in the frame survives that reduction equally badly: your menu bar, your wallpaper, the browser chrome, the notification you forgot to silence, and the one control the video is actually about.
Watching your own recording hides the problem, because you know where to look and your eye goes there. A stranger has no idea. They are reading the whole frame at once, looking for the thing that changed, and by the time they find it you have moved on. That is why so many product demos feel vaguely unsatisfying without being obviously wrong: the information is in there, just not at a size anyone can read.
Doing it by hand, and why nobody keeps it up
In any video editor you can scale the clip up, set a keyframe where the zoom starts, another where it lands, another where it leaves, and repeat. For a two-minute demo with a dozen points of interest, that is somewhere around forty keyframes, each needing a position so the right part of the frame is centred.
It also tends to look wrong on the first attempt. Linear keyframes make the camera move at a constant speed and stop dead, which reads as mechanical. Snapping to each click makes the frame jump around like a bad security camera. Getting it to feel smooth means easing every move by hand, and that is the part people quietly give up on halfway through the second video.
What a zoom has to do to look deliberate
The useful observation is that you already told the editor where to look while you were recording: your cursor went there. A click is the strongest signal of all, because clicking is what people do at the exact moment something matters. Lingering in one spot without clicking is a weaker version of the same thing.
From that, four rules do most of the work.
Arrive slightly early. If the camera starts moving when the click lands, the viewer sees the move instead of the click. Starting about a third of a second before means the frame is already settling as the thing happens.
Ease in and out. Roughly half a second to move in, a little longer to come back out, on a curve that starts and ends at zero speed. A move that accelerates and decelerates reads as intentional; one at constant speed reads as a glitch.
Hold, then merge. Stay in for around two seconds after a click, and if the next click arrives before that hold expires, do not zoom out and back in. Treat them as one moment. Pumping in and out is more distracting than never zooming at all.
Follow loosely, and stay inside the frame. While zoomed in, the camera should track the pointer through a low-pass filter, so a shaky hand does not shake the video, and the crop has to be clamped so it never runs off the edge of the recording and shows black bars.
Doing it with ZonaReq Recorder
ZonaReq Recorder is a Chrome extension that records a tab, a window or the whole screen and does the above for you. While you record it notes where the pointer is and where you click, and nothing else about the page.
- Install it from the Chrome Web Store and pin it, so the icon is there when you need to stop a recording.
- Open the recorder, choose a tab, a window or the full screen, and turn on the microphone if you are narrating.
- Press start. Chrome asks which surface to share, and asks once for permission to follow the pointer. Then work through your demo at a normal pace, and pause for a beat at the things you want the camera to settle on.
- Stop from the extension icon. The editor opens with the recording loaded.
- In the auto-zoom panel, pick an intensity and scrub the preview. Light, normal and close are the same camera path at different strengths, so you are choosing how aggressive it is, not redoing the work.
- Trim the dead air at both ends and export. The file is MP4 or WebM, produced in your browser; nothing is uploaded.
The export runs in real time, because the video is re-recorded frame by frame as the camera moves. A two-minute clip takes two minutes. You can switch tabs while it runs.
Recording the sound that goes with it
A demo with narration needs two sources that behave differently: the app’s own sound and your voice. Chrome can hand the extension both, and they are mixed into a single track as it records.
If the demo is of something outside the browser, that is where the ZonaReq app comes in instead. It captures system audio at the driver level, which is the one thing QuickTime cannot do, and records the screen with the system sound and your microphone kept apart until you mix them.
What this doesn’t fix
Auto zoom makes a clear demo easier to watch. It does not rescue one with no script: if the recording wanders, the camera wanders with it, and all the zoom does is make the wandering legible. Plan the three things you want to show before you press record.
It also only follows what the pointer does, so keyboard-driven work in a terminal gets nothing useful out of it. The free tier of the extension records up to three minutes per take, which is longer than most demos that anyone finishes watching, and a ZonaReq Pro licence lifts the limit on both the extension and the Mac app.