TrailBlaze

Tech Stack, check out our demo videos as well in the comments
User dashboard
Select your pre-made preferences or generate a trip with an AI prompt!
Chatbot (Open AI + Google Maps + Google Places APIs)
Don't like the trip? You can always generate again!
Debug, behind the scenes: using the Google, OpenAI APIs to verify a valid route! It checks opening/closing times and compares prices!
Check out the route! Time/distance, $$$, transport, locations, & Claude-generated summary included
Step-by-step guide, integrated with Maps directions
New to the Big Apple? Take this tour
Image analysis w/ Claude API - this one doesn't show a landmark :(
Image analysis w/ Claude API (Landmark detected!)
Claude AI detected the Empire State Building at night from Bobst 6th floor!
Once you arrive to a spot, take a photo, and upload it to the database!
Social media-like features, likes, comments, reviews! You can copy someone else's route!
FUTURE INTEGRATION: Example of instagram post of what our social media outreach can look like!

Inspiration

Exploring new places should be more than just sightseeing—it should be an adventure. We wanted to create a fun, interactive way for people to experience new locations through challenges, exploration, and discovery.

What it does

Trailblazer is the Strava for travel, turning trips into AI-powered scavenger hunts. Users receive location-based challenges, navigate smart routes, complete tasks with photo verification, and uncover hidden gems along the way.

How we built it

Backend: Node.js for database connectivity and Python for AI-driven chatbot interactions

Database: MongoDB to store itineraries, ratings, and challenge data, Supabase for image data

APIs: Google Maps, Places, and Directions for navigation and location-based challenges, Open AI, Claude API

Frontend: React Native for an intuitive and engaging user experience, Expo Go for fast iOS deployment

Challenges we ran into

Dynamically generating personalized scavenger hunts with real-time API data

Ensuring AI-generated routes were engaging and conflict-free

Seamlessly integrating photo verification and challenge tracking

Accomplishments that we're proud of

Successfully merging AI with travel gamification

Building an interactive experience that encourages exploration

Creating a smooth and dynamic user interface

What we learned

Optimizing API calls for real-time location tracking

Balancing user customization with AI-driven recommendations

The power of gamification in enhancing travel experiences

What's next for Trailblazer

Introducing AR-based challenges for an immersive experience

Social features for competing and sharing adventures with friends

Partnering with tourism boards and brands for exclusive scavenger hunts

Built With

Submitted to

Startup Week Buildathon
- Winner TRACK: The Future of Travel Technology
- Winner Best iOS App

Created by

Even with all our unique features, we needed a solid chat and event-planning system at the core. Using OpenAI for search and idea generation, our software builds a custom itinerary for your day or week, tailored to your accessibility needs, budget, distance, interests, and more.

We then run these plans through Google Maps APIs to check if everything works smoothly. If there's a problem (like not having enough time to get from dinner to a Broadway show) we send it back to OpenAI with the problems, for fixes and better suggestions.

Once everything checks out, the final plan is sent to you, along with notes on any possible tweaks you might want to make.

It all happens in real time, using live data on events happening near you today!

Gregory Butcher
IOS Front-end demo: https://www.youtube.com/shorts/pvAVXhU2kdQ

Erkin Sultonov
We couldn't submit multiple videos - here's a demo of our image recognition and uploading: https://youtube.com/shorts/LctZoEFg0Ag?feature=share thru our app, we have essentially created a public database of images over time for landmarks, this feature is one of the $$$ makers because of image data we store in Supabase along with the location & user route data we store in MongoDB, which can be sold to companies

Alan Wu
CS/DS @ NYU
Abid Ali