PrimaVista: AI Urban Redevelopment Workbench

Early-stage urban redevelopment is notoriously slow, costly, and heavily constrained by zoning laws, heritage rules, and spatial visualization gaps. PrimaVista solves this by providing a collaborative AI copilot for redevelopment companies. Instead of navigating static files and manual databases, teams get a virtual spatial partner directly in their deal room. By inputting an address and a rough redevelopment ambition, the copilot immediately coordinates specialized agents to analyze zoning codes, extract heritage constraints, and build live spatial designs. The copilot proves its enterprise capability by navigating complex, real-world case studies like Milan’s historic Autorimessa Traversi. For this iconic 1930s Rationalist garage—famed for its ramp-less car-lift architecture—the copilot successfully parsed heritage protection laws, modeled structural constraints, mapped 2D interventions, and created interactive 3D volumetric massing models. Technologically, PrimaVista utilizes a robust frontend built with React, Leaflet, and Three.js for interactive spatial visualizations, backed by a Bun and TypeScript server. Its agentic architecture is fully compliant with the swappable JSON-RPC Agent Client Protocol (ACP), allowing developers to easily plug in any LLM runtime for seamless, production-ready planning.