Buckeye-Connect

Our inspiration came from a simple observation. Ohio State is one of the largest universities in the country, yet students often struggle to build real social connections. Clubs do not cover every niche, many interests are too specific, and most friendships happen randomly. We wanted to create a platform that gives students an intentional way to discover people and micro communities that match their interests, whether those interests are academic, athletic, creative, or extremely niche.

Blendr is the result. It is a social discovery platform where students create a profile, list their interests and tags, and then explore recommended people and groups. Every user has a homepage, a profile page, and full access to group discovery, group creation, and joining communities. We wanted the experience to feel familiar and smooth, so we used a Spotify inspired layout to make browsing simple and enjoyable.

We built the initial interface using Lovable, which generated the core UI and navigation. After exporting the project, we extended it in VS Code using React, Next.js, Tailwind CSS, and Supabase. We developed a lightweight matching engine using Node.js. This engine uses tag overlap, keyword extraction, and basic text similarity scoring to generate people and group recommendations. We also added a prompt based search feature where students can type requests such as find students who like anime and data science or show me groups interested in soccer. The system extracts keywords and ranks relevant matches, creating an AI inspired discovery experience.

The biggest challenges we faced were time management, designing a clean user flow, and integrating custom backend logic into a Lovable generated project. It took effort to merge the autogenerated front end with the custom matching system we wrote inside Cursor. Another challenge was keeping the matching engine simple enough to implement within the time limit while still producing meaningful results.

Through this project we learned how to rapidly prototype a fully functional product using modern tools. We learned how to work with Lovable to generate a UI, how to extend it with custom logic, and how to design a text based recommendation system that runs efficiently. We also learned how important clear team coordination is, especially when working on a large idea in a very short timeframe.

Blendr is a prototype, but it shows how a recommendation driven platform can make a massive campus feel smaller, friendlier, and more connected.

Built With

• lovable
• next.js
• react
• supabase
• tailwindcss
• typescript