CampusQuest

Inspiration-

CampusQuest started with a problem close to us: one teammate commutes to Georgia State, while another knows how easy it is to spend too much time inside. Campus can be full of people and still feel isolating. We wanted to give students a reason to look up, explore, learn something, and start a conversation.

We took inspiration from the way location-based games turn an ordinary walk into a discovery. Our twist was to make the discoveries come from the student’s own surroundings—and make the cards they collect a reason to connect with other students.

What it does-

CampusQuest lets students use their phone camera to photograph something around them. AI recognizes the subject and turns the discovery into a collectible card with a name and a short fact or observation. The card reveal makes that moment feel like a reward, and students can save cards to a persistent deck.

The social layer gives those discoveries a purpose beyond collecting. Students can build wishlists, see when someone wants a card they own, and start a conversation about it. The larger vision includes campus quests and in-person card trading: reasons to explore together rather than play alone.

How we built it-

We built CampusQuest as a mobile-first web app using Next.js and TypeScript, with a camera experience that works in the browser and an upload option for testing without a phone. We used xAI to analyze photographed objects and explored image editing to turn discoveries into more distinctive illustrated card art. MongoDB Atlas stores the data behind students’ decks and social features.

We developed in VS Code and GitHub Codespaces, using an AI coding agent to move quickly. The important part was not just prompting it to “build an app.” We repeatedly tested what it made, compared it with our intended experience, and gave precise feedback—especially when a scan looked right but a card would not save, or when the artwork looked exciting but depicted the wrong object.

Challenges we ran into-

Connecting MongoDB Atlas took more troubleshooting than we expected. We worked through an incomplete connection string, network access from our Codespace, and database-user authentication before saved cards would load reliably.

We also learned that a convincing screen is not the same as a working product. At one point, AI recognized a photo and displayed a card, but the save action failed because the recognition could not be verified by the server. At another, an illustrated apple appeared even though it was not what we photographed. Those problems pushed us to focus on the full path from real scan → correct subject → card reveal → confirmed save, rather than judging a feature by its animation alone.

Design was its own challenge. Early cards felt generic, and the first “swipe reveal” was effectively a pre-drawn line. We kept refining the interaction and art direction because collecting a card should feel personal and fun—not like receiving another static AI-generated tile.

Accomplishments that we're proud of-

We turned an ambitious idea into a working camera-to-card experience and got MongoDB-backed saving working after a difficult setup. We are especially proud that we kept pushing on the details: whether the card matched the actual scan, whether it remained in the deck, and whether earning it felt memorable.

We also built CampusQuest around a simple social insight: a card on someone else’s wishlist can be the beginning of a conversation. That connects the game mechanic back to the reason we started—helping students find small, natural opportunities to meet.

What we learned-

We learned how the pieces of an AI product fit together: camera capture, image recognition, generated visuals, server-side verification, a database, and the interface students actually touch. We learned to separate those steps so an attractive result does not hide a failed save or an inaccurate image.

Most importantly, we learned that building with AI still requires strong product judgment. The best improvements came when we stopped asking for a generic “cooler design” and specified exactly what a student should see, do, and feel at each moment.

What's next for CampusQuest-

Next, we want to make campus exploration more collaborative: verified discovery spots, student-created walking routes, group quests, and QR-assisted trades that both students confirm in person. We also want to keep improving the card art so each illustration feels unmistakably connected to the object the student photographed.

Our goal is not to claim an app can solve isolation on its own. It is to create one more reason for students to get outside, discover their campus, and talk to someone new.