AI Square Information

What could AI understand about a facility if it could inspect the space, read its sensors and test proposed changes? AI Square builds Valkyrie, a spatial runtime that brings geometry, equipment, live data and programmable behavior into a shared environment for people and AI agents. A temperature reading tells you one thing. Its relationship to a room, a cooling system, nearby equipment and changing operating conditions helps you ask a better question. Valkyrie gives models that spatial context as well as tools to work with it. With Valkyrie, teams can: • Plan and simulate spaces. Import CAD models or ask agents to assemble environments from 3D primitives. Explore layouts, compare scenarios and review changes before committing to a design. • Understand and operate sites. Connect existing sensors and operational systems to the equipment and locations they describe. Investigate conditions across sources and create rules that combine spatial relationships with live inputs. • Build interactive experiences. Create interfaces, spatial applications and prototypes within the same environment. AI agents access Valkyrie through MCP tools to inspect scenes, modify entities and build behavior. Persistent world state keeps the environment available across agent sessions. People review proposed changes, test results and operational decisions. Applications include airport incursion monitoring and data-center planning and operations. Valkyrie complements existing infrastructure, including building management systems rather than requiring organizations to replace it. Built on AI Square’s own C++ runtime and 3D engine, Valkyrie supports native applications on macOS, Windows and iOS. We’re interested in working with teams building AI agents, industrial applications and the next generation of spatial tools.