CommonPlate: The Future of Social Dining Discovery
"I don't know, where do you want to eat?"
We've all been there. The endless group chat scroll, the indecision, the "anything is fine" (when it really isn't).
CommonPlate solves the social coordination problem of local discovery. It is a real-time, multiplayer dining decision engine that turns the chore of choosing a restaurant into a fun, democratic game. By leveraging Yelp's powerful data, we help groups discover their perfect "plate" in seconds, not hours.
The Problem
Local discovery is currently a solitary experience. You search Yelp on your phone, I search on mine. We send links back and forth. Information is fragmented, and preferences (Dietary, Budget, Vibe) are hard to overlay across a group of 5 people. Who typically makes the final call? The "Loudest Voice in the Room"—not necessarily the best option for the group.
The Solution
CommonPlate is a collaborative discovery layer built on top of Yelp. We democratize the decision-making process.
- Lobby Up: One user starts a session and shares a unique code. Everyone joins the same virtual room via their own device.
- Smart Filtering: The group sets parameters together—Cuisine (Italian vs. Thai), Distance, Price ($ vs $$$), and "Vibe".
- Swipe to Decide: Users are presented with a curated deck of Yelp listings. We use a "Tinder-style" card stack for rapid decision-making. Swipe Right to Yums, Left to Yucks.
- Instant Match: Our algorithm processes votes in real-time. As soon as a consensus is reached (or the deck is finished), the "Winner" is revealed with all the details needed to book.
Powered by Yelp AI
We don't just show businesses; we curate experiences. By using the Yelp AI API as our primary data source, we ensure:
- Trust: Users see real ratings, photos, and review counts they trust implicitly.
- Accuracy: Real-time open/close status and location data to avoid "Heartbreak" (showing a closed spot).
- Depth: Rich attributes allow us to filter for specific needs (e.g., "Good for Groups", "Outdoor Seating", "Vegetarian Friendly").
Future Vision: We are integrating Yelp's new AI Agentic capabilities to not just match, but autonomously call and reserve the table for the winning restaurant, closing the loop completely.
Design Excellence
We believe a hackathon project shouldn't look like a hackathon project. CommonPlate features a "Organic Modern" aesthetic designed to feel premium and calming.
- Mobile-First: Built from the ground up for the "Thumb Zone". Every interaction is reachable with one hand.
- Fluid Animations: Powered by Framer Motion, the card stack physics feel natural and satisfying.
- Visual Polish: Glassmorphism, a curated "Warm Stone" & "Zest Orange" palette, and modern typography (Inter). We focused heavily on the "Tactile" feel of the app.
Technological Implementation
We built a robust, real-time Single Page Application (SPA) using the bleeding edge of the modern web stack.
Core Stack
- Frontend: React 19, TypeScript, Vite (for lightning-fast HMR)
- Styling: Tailwind CSS, Radix UI (Headless accessibility), Lucide Icons
- Real-time Engine: Firebase Realtime Database (for instant session syncing across devices)
- Map Integration: Leaflet for location visualization
Challenges & Innovation
One of the biggest technical hurdles was Real-time Consensus. With multiple users swiping through their own card decks at different speeds, we needed a robust system to track votes in real-time. We implemented Firebase listeners to instantly calculate matches and trigger the specific 'Winner' state for everyone simultaneously, ensuring a seamless multi-user experience.
Potential Impact
- For Users: Eliminates "decision fatigue" and social friction. It brings the joy back to going out.
- For Businesses: Exposes competitive local businesses to groups who might have otherwise defaulted to big chains out of laziness. It levels the playing field for discovery.
- Scalability: The "Lobby + Swipe" model can expand beyond food to Nightlife, Travel Itineraries, and Home Services (contractor selection).
What's Next?
We plan to introduce "AI Chef", a conversational agent that jumps into the lobby chat to suggest hidden gems based on the group's aggregate past history and specific dietary constraints, powered by Yelp's rich business attributes.
Built with love for the Yelp AI API Hackathon.
Built With
- amazon-web-services
- aws-amplify
- aws-apprunner
- aws-ecr
- fastapi
- firebase
- firestore
- gemini
- react
- yelp

Log in or sign up for Devpost to join the conversation.