Odyssey

Inspiration

We wanted to eliminate the chaos of trip planning switching between maps, event sites, and traffic apps. Most itinerary tools are static and ignore real-time context like congestion or local happenings. Odyssey was born to make travel effortless a system that thinks for you, adjusting plans dynamically based on real-world conditions.

What it does

Odyssey is an AI-powered day-trip planner that automatically designs your perfect itinerary. It analyzes your preferred spots, checks live traffic and incidents, and discovers nearby events, crafting the most time-efficient and context-aware route. The result? A personalized, optimized plan that evolves with the world around you your AI co-pilot for perfect days.

How we built it

We built Odyssey as a fully serverless system on AWS: AWS Cognito for user authentication and secure sessions AWS Lambda microservices for each phase (initiation, classification, ETA, and planning) AWS Location Service for real-time geocoding and routing AWS Bedrock (Claude model) for LLM-based reasoning to reorder stops and infer ideal times DynamoDB as a trip-state database to store progress and results A Node.js scraper to pull live local events A clean React frontend to visualize the optimized itinerary and timeline All components communicate asynchronously through API Gateway and DynamoDB streams, ensuring scalability and modularity.

Challenges we ran into

Managing asynchronous data flow between multiple AWS Lambda functions Designing an LLM prompt robust enough to interpret diverse trip contexts and output valid JSON Integrating live traffic and incident data while keeping response times under 2 seconds Handling multi-stop optimization logic that balances distance, event timing, and user preferences

Accomplishments that we're proud of

Fully functional end-to-end trip planner pipeline using real-time data Seamless integration of multiple AWS services into a unified, scalable workflow Successfully orchestrating AI reasoning + geospatial computation for live trip optimization Clean, user-friendly UI that visualizes the entire journey clearly

What we learned

We deepened our understanding of serverless architecture design, geospatial data processing, and prompt engineering for LLM reasoning. We also learned how to efficiently combine event-driven architecture with real-time APIs for production-ready scalability.

What's next for Odyssey

Expand to multi-day travel planning with hotel and meal integration Add real-time re-planning when traffic or events change mid-trip Integrate voice-based itinerary updates through Bedrock agents Launch a mobile version with offline caching and map overlays