BridgeAI

Inspiration

We live in the most connected era in history, yet 1 in 3 Americans do not know their neighbor's name. The inspiration for BridgeAI came from observing a "double crisis" in our local communities: The Economic Barrier: Low-income households often cannot afford essential services (plumbing, tech support, tutoring) despite having valuable skills of their own to offer. The Language Barrier: Non-English speakers are frequently isolated from community resources simply because they cannot communicate their needs.

What it does

BridgeAI is a hyper-local community operating system powered by Gemini Pro. It creates a "Dignity Economy" through three core features: 🤖 Contextual Smart-Matching: unlike traditional classifieds that match keywords, our AI analyzes the intent and urgency of a user's need. It connects neighbors based on complementary skills (e.g., matching a "Need: Tax Help" with a "Skill: Accounting"). 🌐 Real-Time Language Bridges: The built-in chat system removes language barriers. A user can type in Spanish, and their neighbor receives the message in English (and vice versa), enabling cross-cultural support. 🚨 AI-Routed Emergency Response: In a crisis (medical event, power outage), users can trigger an SOS. The AI doesn't just blast a notification; it intelligently routes the alert to the nearest neighbors with specific registered skills (CPR, generators, medical training) for immediate response.

How we built it

Architected BridgeAI as a high-performance, real-time application. Frontend: Built with Next.js and Tailwind CSS for a responsive, accessible interface. Backend: A robust Node.js and Express environment managing a MongoDB database. Real-Time Layer: We utilized Socket.io to power the instant messaging and the sub-second latency required for the Emergency Broadcast system. The Brain (AI): We leveraged the Gemini Pro API. We engineered complex system prompts to handle "Compatibility Modeling"—ensuring matches are safe and equitable.

Challenges we ran into

The "Dignity" Prompt: It was difficult to get the LLM to suggest trades that didn't feel transactional. We spent hours refining the system prompts to ensure the AI suggested "mutual aid" rather than "unpaid labor." Socket Latency: Syncing the translation API with Socket.io was tricky. We had to implement an optimistic UI update so the user didn't feel the lag while the AI was translating the text in the background. Safety Logic: Designing the emergency routing logic was complex—we had to ensure we didn't spam the whole neighborhood, but only alerted those who could actually help.

Accomplishments that we're proud of

Social Impact Potential: We built something that actually addresses the "Equity" theme of the hackathon directly, rather than just forcing AI into a generic app. Seamless Translation: Seeing two devices side-by-side, chatting in different languages with zero friction, was a magical moment.

What we learned

AI as Infrastructure: We learned that AI isn't just for generating text; it is excellent at acting as a "logic router" for human connection. Community Engineering: We realized that technical architecture dictates social architecture. By removing money from the equation, we changed the nature of the interaction.

What's next for Untitled

Voice-First Interface: Adding voice commands for visually impaired and elderly users. ID Verification: Integrating government ID APIs to boost trust scores. Local Gov API: Connecting the app directly to city services for issues that neighbors can't fix (e.g., downed power lines).

Built With

• ai/ml:-google-gemini
• artificial-intelligence-frontend:-react
• express.js
• generative-ai
• git
• javascript
• llm
• mongodb
• next.js
• rest-api-tools-&-deployment:-vercel
• socket.io
• tailwind-css
• typescript-backend:-node.js