Ani

**Uncongest — Less stress between A and B** Navigation apps tell us how to reach a destination, but they don't always help with the decisions that happen along the way: when to stop, where to eat, whether facilities are accessible, or how to accommodate a family with changing needs. Uncongest is an agentic road-trip companion powered by Meta Muse. It anticipates useful stopping opportunities and adapts recommendations as passenger needs, route progress, and arrival conditions change. The project uses two cooperating AI agents: **Agent 1 — Context Understanding:** Interprets natural-language passenger requests, dietary preferences, urgency, and corrections, handing structured context updates to a shared state engine. **Agent 2 — Travel Intelligence:** Operates across two planning horizons: - **Tactical reasoning** evaluates immediate needs and recommends reachable stopping options. - **Strategic reasoning** looks farther ahead at meal and rest opportunities, service gaps, weather, and destination conditions. A deterministic validation layer checks candidate reachability, detours, and available evidence before recommendations reach the driver. Unlike a simple chatbot or place-search wrapper, Uncongest maintains evolving trip context and closes the loop between decisions and actions. Travelers can select a stop, simulate arriving and completing a break, and see their ETA and future needs update. Route changes can trigger fresh planning while relevant passenger preferences persist. The interactive web simulator demonstrates this on real geographic routes, using recorded or live place evidence and clearly identified mock fallbacks. The goal is not to replace turn-by-turn navigation, but to provide an intelligent layer for the human decisions navigation leaves unanswered.