FAQ
Short answers to the questions that come up first.
Does refining re-record my app?
No. The raw capture is kept in the take's working directory, and every visual change is a re-render over it. Your app is driven exactly once per shoot — which matters when the demo involved a real deploy you can't repeat.
Can it demo a logged-in app?
Yes. open-take auth <name> opens a headed Chrome on a persistent profile;
you log in by hand once, then make --profile <name> reuses the session —
still headless. Credentials never touch a plan. See auth.
Why npm create, not npm i -g?
npm create open-take@latest installs the CLI and the agent skill into
the project. A global install gives you the CLI with no skill — and the skill
is the editorial half: it is what makes the demo worth posting.
What does it download?
On the first demo, open-take downloads and caches its own Chrome for Testing
under ~/.open-take/browsers. Later runs reuse it. Nothing else; nothing is
uploaded — capture, render and the editor all run locally.
The demo is good except one moment. Do I start over?
No. open-take beats gives you the numbered map, then one note —
beat 3: no zoom — and a re-render fixes that moment. If it's a taste
question, ab renders the candidates side by side so you can pick by eye.
Can I edit the video in a normal editor?
The mp4 is a normal mp4 — cut it, score it, post it. But the cinematic layer
(zooms, cursor, look) lives in composition.json, so changes at that level
are cheaper as a re-render than as a video edit.
License?
MIT. The repository is github.com/pascal910107/open-take.