01 · Record
Record the way you already work.
MeCorder captures the screen, and beside the pixels it keeps the cursor, every click, scroll and shortcut, so nothing is burned in.
In development · macOS
Don't edit every frame. Record once, then tell the video how to behave when something happens inside it.
macOS 14 or later · Apple Silicon
00 · The idea
A screen recorder sees pixels. MeCorder sees what you did: the button you pressed, the field you typed into, the moment it happened. So instead of editing every frame, you say what the video should do whenever it happens, and the edit writes itself.
The old way
A three-minute demo can have dozens of clicks. Giving each one its own zoom means placing, timing and easing every keyframe by hand, and doing it again when the recording changes.
01 · Record
MeCorder captures the screen, and beside the pixels it keeps the cursor, every click, scroll and shortcut, so nothing is burned in.
02 · Understand
Through macOS accessibility, a click is recorded as the element it landed on: a button called Create Project, a search field, the sidebar.
03 · Program
When Create Project is clicked: focus on it, wait a beat, highlight it, then reset. One Program, applied to every matching moment.
04 · Polish
The Program compiles to camera moves and effects on the Timeline. Preview it, adjust any keyframe by hand, undo it in one step.
05 · Export
What you previewed is exactly what exports, frame for frame. MP4, MOV or GIF, for YouTube, a social feed or a quick message.
Semantic targets
A screen recorder sees a click at a pixel. MeCorder also records what was there, so a Program can say Primary Button and keep meaning it when the window moves or the resolution changes.
Names come from macOS accessibility while recording. If an app exposes nothing, the click has no name and a Program skips it rather than guessing.
Programmable video
A Program says when something happens in the recording and what the video should do about it. It runs on every matching moment, and compiles to camera moves and effects you can still edit by hand.
Every block here works in current development builds of MeCorder. Programs are visual blocks in the app; nothing to type.
Pick an element, and the demo builds a Program for it. Then change the trigger, swap blocks, retime the wait. The video is regenerated from your Program every time.
Pick a target to start.
Choose an element in the recording, or pick one here:
The video on the left is drawn from this object alone: matched against the recording's events, laid out in time, and compiled to camera keyframes and effects, the same way MeCorder does it.
Beyond single rules, Programs have the parts you'd expect from a real language, shaped as blocks.
One Program can tell clicks apart: IF the target matches, do one thing, ELSE another.
Start several things at the same moment, then carry on when every branch has finished.
Count, remember and compare. Later events see what earlier ones changed.
Turn a run of blocks into a reusable Behaviour with parameters. Edit it once; every Program follows.
Scenes from the Timeline get ENTER and EXIT blocks, shots that start on an event, and Cut or Fade transitions.
✦ “Make the camera follow every button click”
Describe what should happen in plain words. The proposal is checked against the recording and shown as blocks before anything changes.
Loops, an IF directly inside another IF, variables shared across a whole project, and finding elements visually in apps that expose no accessibility information are not built yet.
Two views, one edit
Timeline: edit what happened. Clips, scenes, camera keyframes, text, effects and audio, laid out in time and adjustable to the frame.
Program: define what should happen. The rule behind those keyframes, written once and applied to every moment that matches.
Program doesn't replace the Timeline. It writes to it: every block becomes ordinary keyframes and effects that you can open, move or delete.
Timeline: a click at 4.3 seconds; a camera move to the Primary Button; a 0.3 second hold; a highlight on the Effects track; the camera resets. Program: When Click Primary Button, Focus Primary Button, Wait 0.3 seconds, Highlight Primary Button, Reset camera.
Before and after
Drag across. On the left, what the screen saw. On the right, one Program for every click: the camera finds each element, the cursor gets weight, every press lands.
Everything else, done properly
Zooms are generated from what you did the moment you stop recording, with styles from Calm to Lively. Every move is a keyframe you can retime, re-aim or ease differently.
The cursor is recorded separately from the pixels, so you can resize it, smooth its path, change its look on click and drag, or hide it, after the fact.
Scenes, clips, camera moves, clicks, keys, text, effects, webcam and audio on one Timeline. Cuts never touch the original, so anything can come back.
Tell it where the video is going: YouTube, a vertical feed, a message under 25 MB, or a GIF. The file is read back and checked before it replaces anything.
Record once. Program the behaviour. Export a video that feels designed.