Skip to content
Indoor mapping

From CAD upload to navigation mesh, before your coffee is ready.

MindDigitize ingests the file format your architect already uses. Walls snap. Rooms get inferred. A pathfinding graph is generated. You don't see any of it — you see a map you can route on.

  • DXF · DWG · IFCsupported on day one
  • 8 minmedian time to first map
  • 1.2Mline segments / file ceiling
FLOOR-02.dxfINFERRED
What the parser does for you

The pipeline isn't the product. The map is.

  • Wall snapping

    Stray endpoints get welded to their nearest neighbours within a configurable tolerance. The parsing report tells you exactly where we cleaned up.

  • Room inference

    Closed loops become rooms. Open corridors become navigable space. Doors are detected as breaks in wall continuity.

  • Text-block harvesting

    Floor-plan annotations become POI candidates with their original coordinates preserved. You decide which ones to publish.

  • Topology graph

    A skeleton graph (Zhang-Suen thinned + simplified) gives the router something to walk before you've placed a single waypoint by hand.

How it looks from the API

Upload a file. Get back a map.

route.ts
// POST a DXF, get a parsed Map back
const res = await fetch('https://api.minddigitize.com/v1/maps', {
  method: 'POST',
  headers: { Authorization: `Bearer ${token}` },
  body: form // multipart with the .dxf
});

const { mapId, floors, rooms, issues } = await res.json();
// floors: [{ id, level, rooms: 18, area: 4210 }, ...]
// issues: [{ kind: 'open-wall', floor: 'L2', point: [x,y] }, ...]
What you don't have to do

The work nobody else takes off your plate.

  • Hand-tracing rooms over a PDF.
  • Wrestling with rogue CAD layers and exploded blocks.
  • Re-routing every time facilities moves a wall.
  • Maintaining three exports for three vendors.
Ready to ship

Your building is one CAD file away from a navigable map.

Book a 20-minute demo and we'll have a working public viewer of your floor plan before the call ends.