On a Meta Quest 3, open 0xr.space in Quest Browser or the free Model Explorer app, open any model, enter the room and press Scan on the dock. Things the scan recognises within about 3.5 m get a soft box and a label with their name and distance, such as cup · 45 cm, and the label you look at shows the thing's size. Press a thing's name, or say it to Zero: the other boxes go, a panel shows Making 3D model with a percentage, and a photo of the thing goes to Photo → 3D, which builds a textured model. Opened within ten minutes, the model is scaled to the size the scan measured. While a scan runs the control reads Stop; pressing it ends the scan and builds nothing.
Scan is experimental and works on Quest 3 only, not Quest 3S. The far side of the model is guessed from one photo. Good light helps; in a dark room the scan stops and says Too dark. Turn on a light. Scan is not available in every country.
Not with boxes yet. On a Quest 3S a press reads a QR code if one is in view and otherwise makes one photo into a model with Photo → 3D.
The browser was given a virtual camera. In Settings, allow Headset cameras for Browser, then press Scan again.
See the privacy page at https://0xr.space/privacy/ for what the camera features send and keep.
Everyday objects such as cups, bottles, bowls, phones, laptops, keyboards, books, vases, plants and fruit, and bigger ones such as chairs, sofas, tables and TVs, which are drawn as an outline. A name Scan does not know, said to Zero, is looked for by name.
No. It is built from one photo, so the back is guessed. The size comes from what the headset measured, when it measured one.