Environment Memory

Environment Memory is a mobile web prototype that gives your workspace a visual memory. Start a scan, show your objects, and add your own names, groups and uses. The system connects them to dated observations, places and containers in a relationship graph. Later, ask where something was last seen and inspect the image behind the answer. Our demo follows a small robotics workbench: scan the tools, name one “my electronics screwdriver,” associate it with the Robot project and servo mounting screws, then retrieve its recorded place through the mobile interface. AI interprets images and instructions; application code validates updates, preserves evidence and supports correction and undo. A unique saved-name lookup can run without another model call. The fictional workshop enables fast, repeatable development. We also tested the prototype with real camera images, with successful results in those tests. No personal photos or local test records are included in the public source or video. Broader real-world validation remains future work. Similar objects and closed-box relocation can still be confused, so matches are provisional and correctable.