Wander Wise

Your personal travel recommendation app.

Comment

Inspiration

I was inspired to build WanderWise after experiencing frustration with generic travel guides that missed authentic local experiences. The Qloo API hackathon presented the perfect opportunity to explore how specialized AI services could work together. I was particularly excited about Qloo's Taste AI API for its data-driven restaurant recommendations and wanted to see how it could complement GPT-4o-mini's broader travel knowledge to create truly personalized travel planning.

What it does

WanderWise is an AI-powered travel companion that provides personalized recommendations for any destination worldwide. Users enter their destination, trip duration, budget preferences, and personal interests, and the app combines Qloo's specialized restaurant AI with GPT-4o-mini's comprehensive travel knowledge to deliver a complete travel guide. The app provides destination information, must-try restaurants, cultural experiences, hidden gems, and practical travel tips - all tailored to the user's preferences and trip duration.

How we built it

I built WanderWise using a Flask backend with parallel processing architecture where both AI APIs are called simultaneously to minimize response times. The system uses Qloo's Taste AI API to search for locations and retrieve curated restaurant recommendations, while GPT-4o-mini provides comprehensive travel guidance including experiences, hidden gems, and cultural insights. The frontend is built with vanilla JavaScript and CSS for a responsive, modern interface. The application is deployed on Railway for optimal performance and reliability.

Challenges we ran into

The main challenge was dealing with deployment platform limitations - initially using Render's free tier caused timeout issues when the combined API calls exceeded 30 seconds. I solved this by switching to Railway which offers better free-tier performance. Another significant challenge was optimizing response times while maintaining recommendation quality, which I addressed by refining prompts, reducing token usage, and implementing parallel processing. I also had to create robust error handling to ensure the app remains useful even when external APIs fail.

Accomplishments that we're proud of

I'm proud of successfully orchestrating dual AI systems where Qloo's specialized restaurant AI and GPT-4o-mini's general travel knowledge work together synergistically. The application is fully production-ready with sub-30-second response times and comprehensive error handling. The project demonstrates real-world viability of AI-powered travel planning and shows how different AI approaches can complement each other rather than compete. The clean, responsive interface provides an excellent user experience while showcasing the power of AI integration.

What we learned

I learned valuable technical skills including API orchestration for coordinating multiple AI services, Flask development for production applications, and prompt engineering for consistent AI responses. Most importantly, I discovered how specialized AI services can work with general AI to create better results than either could achieve alone. I also gained insights into deployment optimization, error handling strategies, and the importance of graceful degradation when dealing with external APIs.

What's next for Wander Wise

The next steps for WanderWise include expanding the Qloo integration to include bars, cafes, and other venue types, implementing user accounts to save favorite recommendations, adding a database to cache popular destinations, and developing a mobile app version. I also plan to integrate additional AI services for weather predictions, local event recommendations, and real-time translation features. The goal is to create the most comprehensive AI-powered travel planning platform that combines multiple specialized services for an unparalleled travel experience.

Built With

Share this project:

Updates