Spatial Agents
Teach an agent to see, understand, and remember the real world. From walkthrough footage to drone surveys — speak, annotate, and build shared spatial knowledge.
Built for
Voice Native
Voice-first platform
Walk around, talk, and your words are pinned to the 3D space. Audio is extracted from video and anchored to real-world coordinates — every observation lives exactly where you said it.
“This beam doesn't look right. Can you pull up the structural plans and check if this matches the spec?”
Walk a job site scan, narrate what you see. The agent cross-references plans and pins findings to the exact 3D location for the next crew.
“I wonder what this would look like with a marble countertop? Show me Calacatta, honed finish.”
Think out loud while exploring a space. The agent captures your ideas and pins material notes exactly where they belong.
“Enemy observation post at this ridgeline. Two vehicles, last movement 0930. Mark as priority target.”
Pin battlefield intelligence from drone footage directly into the 3D reconstruction. Brief the next rotation instantly.
“Structural collapse here, possible victims underneath. Mark as high priority search area.”
Build situation awareness from aerial scans. Every responder sees the same annotated 3D picture.
Spatial Indexing
Pin information
in 3D space
Every note, observation, and measurement is anchored to real-world coordinates. Search, filter, and navigate your spatial knowledge base.
- Deploy virtual AI drones to autonomously index and catalog your space
- Anchor notes to exact 3D coordinates
- Search across all spatial annotations
- Share annotated scenes with your entire team
Video to 3D
One video.
One 3D world.
Upload a single video and Rotar reconstructs a full 3D Gaussian splat scene. Walk through it, annotate it, share it — as if you were there.
Collaboration
Share with anyone
One link. Full 3D scene with all annotations. No installs, no plugins, no friction.
Start building with
Spatial Agents
Turn any video into an interactive, annotated 3D world.