WILDFOLD — A Living Field Study

A small world. A living balance. WILDFOLD is a local ecosystem sandbox and restoration game. Build a patch of woodland, wetland and meadow, then observe six kinds of animals using it. Each simulated individual has needs, species parameters, local perception, remembered resources, routes and a reason for its current action. Observe European rabbits, roe deer, red foxes, wild boar, mallards and grey herons with distinct low-poly models and articulated movement. Click a life to see its decision, food, hydration, rest, health and remembered resources. Follow it through a terrain-aware 3D habitat, or hide the interface to watch the community. Restore meadow, woodland, scrub, pond, stone and shelter tiles. Introduce appropriate animals. Try six restoration briefs, from providing rabbit cover to maintaining a mixed community through a dry spell. The conditions are evaluated by the simulation and must be sustained for a stated duration. Pause at any point, choose a requested simulation speed from 1x to 8x and switch between mild, dry and wet conditions. Keep a local checkpoint before an intervention. Export a portable .wildfold habitat or import one after validation. A hidden or closed tab does not calculate additional background population losses. The application requests no permissions or host permissions. It does not inspect websites, use accounts, display advertising, track usage or call an online AI service. Game code, graphics and icons are packaged locally. Optional natural-history reference links open only when clicked. Local IndexedDB saving requires browser support; an unavailable-storage session is explicitly labeled temporary and must be exported to keep it. Scope and requirements: WebGL 2 and a desktop Chrome environment are required. There is a fixed 24x24 habitat and a maximum of 48 simulated animals. Time, food, births and behavior are simplified for gameplay. Birds walk or swim rather than fly. Predation is limited to foxes hunting rabbits. Genetics, disease, migration, seasons and scientifically validated population prediction are not provided. Performance depends on the device; requested speed is not a guaranteed real-time rate.