Unified foundation model for promptable segmentation
Semantic search over live-camera snapshots with CLIP
Calibrate a photo to get camera intrinsics and gravity