Turning 3D Scans Into Editable 3D Scenes

I gave AI a 3D scan of my parents’ old house, and it rebuilt the room in Blender. Then I made the furniture float. Workflow: ~400 photos I took in 2019 → RealityScan photogrammetry mesh → Astra rebuilds it as an editable Blender scene, using my photos as reference and the scan textures to make procedural materials. It didn’t download a single asset. Every object is editable, which is why I could just turn off the gravity. Turns out multimodal LLMs might have crossed the chasm on inverse graphics. And I’m still convinced the killer use case for 3D capture is memory. Scan the places that matter to you.