Mobile Physical AI
We pinned the documentation onto the robot — in 3D.
The Petoi Bittle Spatial Twin lets Atelier agents navigate, inspect, verify and intervene in operational environments — continuously synchronizing real-world observations with simulation. Orbit it below, then click the marker floating on it: a walkthrough opens, anchored in 3D space, right where the work happens.
Petoi Bittle — interactive 3D Gaussian splat. Click the marker to open the walkthrough.
A robot you can interrogate
This is a spatial twin of a Petoi Bittle — one of the most accessible ways to get genuinely hands-on with Physical AI, a palm-sized quadruped you can actually program and experiment with. We added the spatial layer: a browser-based twin you can orbit, with context attached in place.
Annotations anchored in space
The markers floating on the twin aren't labels on a slide — they live in 3D, attached to the machine. Click one and a walkthrough video plays. The same anchors can hold a spec sheet, a reference image, or a link. The documentation lives on the thing, in space — not buried in a folder of PDFs.
Why it matters for Physical AI
Onboarding, field service, training — the moments that decide whether a deployment works are usually in front of the machine, not at a desk. Putting the right context exactly where the work happens turns a 3D model into something operational.
It is a small piece of how we think about operational spatial runtimes: the twin isn't just to look at — it's where knowledge, context and the real world meet.