Inspiration
We started from a simple question: “How many of our own neighbors do we actually know?” When we dug into the data on loneliness and the fact that 1 in 3 Americans don’t know their next-door neighbor’s name, it felt wrong that we’re more connected to strangers on apps than to people on our own street. We also noticed that people are much more willing to help when it’s a fair exchange—“I tutor your kid, you give me rides”—instead of one-way charity. That idea of mutual, cross-income help is what inspired Good Neighbor.
What it does
Good Neighbor is a local skill & item sharing platform. Neighbors create a profile with “things I can offer” and “things I need.” Our AI then surfaces matches that are complementary (your math tutoring for their home repair, their extra laptop for your resume help) and intentionally highlights cross-income “bridge” connections. The app suggests friendly first messages, supports messaging between neighbors, and tracks positive activity on a community leaderboard so people get recognition for helping and being helped.
How we built it
We designed the experience around a clean, warm, non-boxy UI and then implemented it with a modern web front end (React-style component architecture with utility-first styling). We built core flows: a landing page, profile creation, a Discover screen with AI-powered filters and a community map, a messaging interface, and a leaderboard. On the backend, we structured data models for users, offers, needs, matches, and messages, and set up endpoints that can plug into embedding models and LLMs for matching and message generation.
Challenges we ran into
Technically, balancing time between a polished front end and meaningful backend/AI structure was tricky, and our first attempt at a fancy 3D map looked cool but hurt usability—so we simplified it into a clearer, friendlier community view.
Accomplishments that we’re proud of
We’re proud that Good Neighbor feels like a real product, not just a slide idea: you can go from landing page → create a profile → discover neighbors → start a conversation → see yourself on a leaderboard. The AI-powered filters and match UX make it obvious why two people should connect, instead of hiding the logic in a black box. We also managed to keep a strong, consistent visual language throughout—warm colors, pill-based chips, and friendly copy that reinforces dignity and reciprocity.
What we learned
We learned how important it is to design AI features around human awkwardness, not just algorithms—“What do I say first?” is as real a problem as “Who should I meet?” We also got practice structuring a project so that front end, backend, and AI pieces line up into one coherent demo story. And we learned that small details—empty states, microcopy, and gentle animations—go a long way toward making a social app feel safe and welcoming.
What’s next for Good Neighbor
Next, we want to plug in real embedding-based matching and LLM message generation, and store real interaction data so we can measure impact. We’d add verification and lightweight reputation, plus AI-clustered community events (like resume workshops or tool swaps) based on real neighborhood needs. Longer term, we’d love to pilot Good Neighbor with a local community center or city, and build dashboards that show how many hours of tutoring, rides, and shared tools a neighborhood is generating each month.
Log in or sign up for Devpost to join the conversation.