Skip to content
/
    how it works

    How it works

    Scan and place an anchor, share the payload any way you like, then find it from an independent scan on any other device. No cloud needed, no markers.

    Scan and place an anchor
    One device scans the room’s mesh geometry (its 3D shape) and places an anchor. A compact payload (anchor-local mesh features and a room fingerprint, no images) builds automatically.
    geometry only, no photos
    Share the anchor file, any way you like
    AirDrop it, email it, or host it behind your own backend. The payload is just a file, typically 2-5 MB. Nothing about sharing it depends on this project at all.
    a 2-5 MB file you own
    Any device finds it from its own scan
    The second device scans the room on its own and opens the file. The algorithm matches that scan to the payload and finds the anchor’s exact spot. No cloud needed.
    independent scan, own frame
    Multi-room batches
    Register anchors across many rooms, then rank and localize a live scan against all of them at once.
    Robust to moved clutter
    Small objects (cups, chairs) moved between scans are tolerated, measured on same-session captures. Heavy clutter displacement can still bend the match; see the pipeline topic.
    Any mesh source
    HoloLens spatial mesh, ARKit/ARCore scans, or depth-only capture: all feed the same geometry-only pipeline.
    The pipeline, technically
    Save, share, find, verify: the algorithm names behind the geometry-only pipeline.
    Offline by design
    No images leave the device, no cloud service, works in the dark on depth-only capture.

    Curious how a match is verified before it’s ever reported? See how it knows it’s right. Ready to try it? Start a free trial or read the quick start.