From a raw capture to something worth shipping.

Product work rarely starts
with a tidy image file

Screenshots, recordings, live pages, and social posts all come in the same door. Whatever you bring becomes the same kind of canvas, so everything after this behaves identically.

Framing decides what it is
styling decides how it lands

Every treatment here is non-destructive — the original capture is kept behind each crop, frame, and filter, so nothing you try is a one-way door. Save a finished look as a preset and re-apply it across a whole set.

A screenshot says here it is
context says here is what matters

Labels, marks, and extra shots are what turn a picture of a screen into an explanation. Each one lives on its own layer, so you can rearrange the argument without re-cropping the evidence.

Motion built by describing states
not by scrubbing keyframes

Each canvas carries its own timeline. The editor eases from the committed values into whatever a clip changed, so you write the destination and it works out the path.

The end of the same pipeline
not a separate render service

Stills are rasterised from the canvas you have been looking at, and video is encoded on your own machine. The file appears without a queue, a watermark negotiation, or an upload.

Opening a capture
does not upload it

Editing, rasterising, and video encoding all run in the browser, and your work in progress is stored on the device, so a reload or a dropped connection does not cost you the composition. The server side is deliberately small and always explicit: capturing a URL, fetching a social post, syncing a cloud draft, publishing a share link. In each case it is the request you made or the rendered output — never the source capture sitting on your canvas. The project is open source under AGPL-3.0, so the boundary can be read rather than trusted; see the about page or the privacy policy.