Film your real interface performing the real steps — clicks, typing, hovers — and re-render when the product changes instead of re-recording.
One release moves a button and your library teaches people the wrong thing.
Nobody minds making the first screencast. They mind making it again every quarter.
Support answers the same question in prose because there is no video to link.
Five to eight of them. If you cannot list them, the tutorial is not ready — and the list becomes the scene structure.
A shot can click a button, type a real query, hover to open a menu, or spotlight one control while the rest dims. The viewer sees it happen.
The film is generated from the live page, so re-rendering captures today's interface and keeps the narration, structure and timing.
Yes, with a capture session you register once from the VS Code or browser extension. It is scoped to that one origin and expires within three days, and it is a Scale-plan capability. Film a demo account, never real customer data.
Yes. A shot performs its action in a real browser before the frame is captured — a click, typed text, a hover, or a spotlight on one element — so the viewer watches the step rather than reading a description of it.
As long as the steps take, usually 60 to 150 seconds. Tutorials are the one format where length is not the enemy: the viewer has already decided to watch and is following along on their own screen.
Paste a URL, watch the AI build it. If it's not the fastest video you've ever produced, it cost you nothing.
Start free →