2026-07-23 · AutoCast
Manual zoom keyframing in a traditional editor is a two-pass job: record the screen, then go back and mark in/out points for every moment you want the viewer’s attention pulled somewhere. For a ten-minute tutorial with forty distinct clicks, that’s forty decisions about when to zoom in, how far, and when to ease back out — done after the fact, disconnected from the moment that made the zoom necessary in the first place.
The expensive part of manual keyframing isn’t setting one keyframe — it’s staying consistent across all of them. Zoom in too fast on click 12 and too slow on click 30, and the video reads as uneven even if no single cut is wrong. Editors solve this by building templates or using consistent easing curves, which is itself setup work.
Cursor-driven auto-zoom collapses that pass entirely: because the zoom is anchored to the click event itself, timing consistency is structural, not something you enforce by eye across forty manual edits. You pick one of three zoom styles — Punchy, Standard, Cinematic — once, and every click in the recording inherits that pacing. What used to be per-clip judgment becomes a single upstream decision.
The tradeoff is real: a human editor deciding shot-by-shot can zoom into something you didn’t click — a status message that appeared passively, a background element worth calling out — and can hold a zoom longer than the interaction itself warrants for narrative effect. Auto-zoom can’t originate attention on something you didn’t interact with, because it has no signal to key off besides your input.
This is also why AutoCast anchors dubbed audio timing to the same clicks and zooms rather than to word count alone — when the visual attention and the audio pacing share one clock, translated audio that runs shorter or longer than the source language doesn’t drift out of sync with what’s on screen.
Override manual zoom for anything auto-zoom structurally can’t see: narration-only stretches with no clicks, a moment where you want to hold on a result after the interaction that produced it, or comparison shots where you’re contrasting two states rather than clicking through one. Auto-zoom handles the mechanical majority of a tutorial’s attention cues; manual adjustment is for the handful of moments where the story diverges from the click stream.
Read more on why dubbing follows clicks and on choosing an auto-zoom style that matches your content’s pacing.
See also: software tutorials