We read behaviour, not content.
mvoom opens pages that belong to other people. That is only defensible if what comes out the other side is a fact about motion and nothing else. So the destruction happens inside the worker — before anything touches disk, and before any model sees it.
Copy is destroyed in memory. What survives is that a text node existed and how it moved.
No pixels are stored. A layer keeps its box and its timing, not its picture.
Never extracted, never redistributed. Metrics only, where geometry depends on them.
Logos, colours as identity, names. None of it is part of a recipe.
Boxes, positions, the shape of what moves
Stacking and grouping, because a stagger needs siblings
Duration, delay, curve, trigger, stagger
mvoom reads one component at a time. Selecting a whole page aborts the read — the page bundle was designed and then retired on purpose. Cloning a site is not a use we support, and the product is built so it is not a use we can be talked into either.
You can ask us not to read it, and the exclusion is honoured across the service.
OPT OUT A SITE →