Inspiration

During a trip to Singapore, I tried to plan my itinerary using ChatGPT. Initially, it seemed like the perfect solution, but I quickly ran into problems. The model was not up-to-date on current rates, bus schedules, or special holidays such as Republic Day. These gaps in information disrupted my entire trip and made me realize the need for a smarter travel planning solution. This experience inspired Voyalytics, a system designed to create personalized itineraries while accounting for real-world constraints and dynamic data.

What it does

Voyalytics allows users to plan trips tailored to their unique needs. It considers budget, time constraints, wheelchair accessibility, and dietary restrictions to craft a complete itinerary. Users can optimize their plans based on weather conditions and view crime analytics to avoid unsafe areas. The system ensures that each trip is fully customized, reliable, and adaptive, giving travelers the confidence to explore new places without uncertainty.

How we built it

We integrated multiple APIs and tools to bring Voyalytics to life. Google Maps API was used for geocoding and directions to ensure accurate routing. Weather API integration allowed itineraries to adjust based on forecasts. ElevenLabs API was employed to provide natural voice narration for the itineraries. At the core, we leveraged a language model to reason about travel preferences and generate structured outputs, which we then parsed into actionable itineraries. This combination of real-time data, personalization, and reasoning created a seamless travel experience.

Challenges we ran into

One of the major challenges was integrating multiple APIs and ensuring they worked together smoothly. Parsing the language model outputs was complex because we had to ensure consistency and prevent hallucinations. Coordinating dynamic information from weather, accessibility, and safety sources to generate reliable itineraries required careful design and validation.

Accomplishments that we're proud of

We built a fully functional application in just 24 hours. Voyalytics successfully aggregates and reasons over information from multiple web sources to create personalized travel plans. The system adapts to individual user needs while providing accurate, actionable, and practical itineraries.

What we learned

We gained hands-on experience with ElevenLabs for text-to-speech narration. We learned to use Google Maps APIs for geocoding and directions. Weather API integration taught us how to dynamically adapt plans based on environmental conditions. Most importantly, we learned techniques to parse and structure language model outputs reliably so that they can be used in real-world applications.

What's next for Voyalytics

Future plans include live-time updates based on news and events, integration of other forms of transport, and expanding beyond New York to cover multiple global destinations. Our vision is to make Voyalytics a fully adaptive travel companion that responds to real-world changes, personal preferences, and safety considerations.

Built With

Share this project:

Updates