Boston Scout

The inspiration for our project came from the realization that travel planning often feels like a fragmented and overwhelming chore. We wanted to transform the urban exploration experience by merging real-time transit data with the authentic local wisdom found in community discussions from reddit. Our goal was to create a tool that did not just list landmarks and locations, but instead understood the heartbeat of the city itself.

We built the application using a high-performance stack of React and TypeScript, supported by a MongoDB backend. The platform integrates Google Places and Grounding with Google Maps for location discovery, the MBTA API for live transit updates, and ElevenLabs for voice navigation. A key differentiator is our use of Gemini AI and a custom Reddit scraper to power a Retrieval-Augmented Generation system. This system offers users deeply contextual insights derived from real human conversations.

We faced significant challenges in design and user-friendliness. Mapping high-density data onto a 3D interactive interface was a delicate balancing act. We wrestled with ensuring the AI Optimizer felt intuitive rather than robotic. Through relentless UI/UX iteration, we successfully turned a complex technical backend into a seamless tool for every traveler.

Built With

• eleven
• elevenlabs
• express.js
• geminiapi
• google-places
• mongodb
• node.js
• react