Skip to Main Content

UMBRA is a lunar exploration simulator built to test a harder question than whether AI can write code: can it understand human intent, act independently, and recognise when to stop? Two GPT-6 Astra agents coordinate as Mission Control and an onboard rover across a simulated Earth–Moon communication delay. A commander gives a plain-English goal—“Find water ice near the shadowed basin without risking the drill”—and approves the mission. Astra chooses where to explore, adapts when terrain becomes hazardous, and refuses missions it cannot safely return from. Built on real lunar elevation and imagery distributed through NASA’s Moon Trek, UMBRA combines calculated illumination with simulated ice deposits and rover operations. The model receives terrain and illumination data, but never the hidden ice map: it must reason about where to search. I built UMBRA to explore what trustworthy autonomy looks like where human intervention cannot arrive immediately. The defining demonstration is not just a rover completing a mission—it is a rover explaining why an unsafe mission should not begin.

Team

More from GPT-6 Astra Hackathon London

We reimagined how you visualise and interact with your agents, by importing sessions from codex, and bringing these agents to life. You can invite other…

dot-dash project preview

Sim Sketch Artist turns your selfie into a portrait drawn by a robot in simulation. Capture a photo through our Lovable app, and GPT-6 Astra transforms it…

Wayfinder helps patients and visitors find their way around hospitals using their phone camera. Hospitals can be difficult to navigate, especially when…

Made to Fit helps electronics hobbyists turn the parts they already own into useful projects. Upload photos, review Astra’s suggestions, confirm measurements…

The Object Knows is an agentic investigation layer for the physical world. Most visual AI stops at recognition: it tells you what it thinks it sees. The…

The Object Knows project preview

Cereal is a live interpreter and translation app for two people who don't share a language, sharing one pair of AirPods. I’ve had the idea and some designs…

@lujstn (Lucas Johnston Kurilov) project preview