TripGenie — Project Story 💡 What Inspired Us Travel planning is often overwhelming—multiple tabs, scattered itineraries, unpredictable weather, overcrowded tourist spots, and the constant fear of plans falling apart. We wanted to create a system that thinks like a human travel expert: adaptive, personal, weather-aware, and capable of instant backup plans. This inspired the creation of TripGenie, an AI-powered smart travel companion. 🛠️ How We Built It We built TripGenie using a React + Vite frontend, a Node.js backend, and Llama API for AI reasoning. Key components include: A clean UI for itinerary, dashboard, Plan B, and nearby explorer. A backend wrapper that embeds logic for weather handling, preference learning, pricing hints, and smart re-planning. Shuffle/Upvote features that adapt preferences over time using a simple weight system:

new weight

old weight + Δ new weight=old weight+Δ A fallback “Plan B Mode” triggered during closures, heatwaves, rain, or crowding. 📚 What We Learned We learned how to structure prompts for consistent AI behavior, implement user preference learning, design weather-aware itineraries, manage full-stack communication, and build a scalable UI. We also gained deeper understanding of API handling, React component architecture, and real-world UX design. 🚧 Challenges We Faced Making the AI location-specific instead of giving generic results Creating weather logic without removing major attractions Building preference learning without a trained model Designing Plan B logic that adapts instantly Debugging deployment issues and integrating multiple services smoothly