Inspiration
Travel planning is often time-consuming and fragmented—users juggle multiple websites, check availability, and compare prices manually. We wanted to create an AI agent that makes travel and hotel booking seamless, personalized, and efficient.
What it does
TripHack autonomously searches for hotels, compares prices, checks availability, and suggests optimal itineraries based on user preferences. It can also integrate real-time data like weather, local events, and travel restrictions to make smarter recommendations.
How we built it
We used AWS Bedrock to host the Large Language Model, AgentCore for autonomous decision-making, and integrated APIs for hotel booking and travel information. The agent leverages reasoning LLMs to understand user preferences and dynamically adjust recommendations. Optional AWS services like Lambda, S3, and API Gateway manage workflow automation and data storage.
Challenges we ran into Integrating multiple hotel booking APIs with different response formats. Ensuring real-time updates for availability and pricing. Designing the AI to handle ambiguous user inputs and preferences.
Accomplishments that we're proud of Successfully built a fully autonomous AI agent capable of planning a trip end-to-end. Developed a system that can reason over multiple factors like price, location, and events to make smart recommendations. Achieved smooth integration of external APIs and real-time data handling.
What we learned Building a reasoning-based AI agent requires careful orchestration of multiple AWS services. LLMs can significantly enhance decision-making but need structured input/output pipelines. Real-world integrations (APIs, databases) often require creative error handling and normalization of data.
What's next for TripHack Add personalized travel suggestions using user travel history. Expand to include flights, car rentals, and activity bookings. Implement a mobile-friendly interface for live chat-based travel assistance.
Log in or sign up for Devpost to join the conversation.