Social Dining

Stop arguing, start eating. A collaborative social dining app helping groups decide where to eat using Yelp AI API. Invite friends, set preferences, and let AI find the perfect restaurant for everyone

Comment 1

Social Dining 🍽️

Stop debating. Start eating.

Inspiration

Anyone who has tried to plan a group dinner knows the pain: endless back-and-forth messages, conflicting diets, different budgets, and no clear winner. One person wants vegan, another wants steak, someone else cares about ambiance — and suddenly dinner becomes a negotiation.

We saw an opportunity to turn this chaos into collaboration. Instead of forcing one person to decide or relying on simple filters, we asked:

“What if AI could act as a neutral mediator — understanding everyone’s preferences, resolving conflicts, and guiding the group to a decision in minutes?”

That question inspired Social Dining, a collaborative, AI-powered experience designed to help groups stop debating and start eating.

What it does

Social Dining is an AI-powered social dining app that helps groups decide where to eat — together.

Here’s how it works:

  • Session Creation A host creates a dining session with a location and time and shares a link with friends.

  • Preference Collection Each participant joins the session and submits their:

    • Dietary needs
    • Cuisine preferences
    • Budget
    • Desired vibe

  • AI Conflict Resolution Before searching, the system analyzes the group profile to detect conflicts (e.g., vegans vs meat lovers) and determines a smart compromise strategy.

  • AI-Curated Recommendations Using the Yelp AI API, the app generates top restaurant picks — each with:

    • Why it was picked
    • Trade-offs and honest downsides

  • Real-Time Voting Participants vote on recommendations, and a live leaderboard shows the current favorite.

  • AI Reservation Agent (Simulation) Once a winner is chosen, the host can ask the AI to “book” the table, simulating a real restaurant call with realistic success or failure outcomes.

In just a few minutes, groups move from indecision to a confident choice.

How we built it

Social Dining is built as a modern, cloud-native application with AI at its core:

  • Frontend Built with Next.js 14, React, and Tailwind CSS, optimized for mobile-first, fast group interactions.

  • Backend A high-performance FastAPI service that orchestrates sessions, preferences, voting, and AI workflows.

  • AI Layer

    • Yelp AI API for semantic restaurant discovery and reasoning
    • Custom AI logic for conflict detection and recommendation strategy
    • A mock Reservation Agent that simulates real-world booking behavior

  • Data Layer Supabase (PostgreSQL) stores sessions, participants, recommendations, and votes with real-time update support.

  • Deployment Fully containerized services deployed on Google Cloud Run.

This architecture allowed us to iterate quickly while keeping the system clean, scalable, and hackathon-ready.

Challenges we ran into

One of the biggest challenges was designing AI behavior that felt helpful and transparent, not mysterious. Simply returning restaurant results wasn’t enough — the AI needed to explain why a place was chosen and clearly communicate trade-offs.

Another challenge was modeling conflict resolution. Group preferences often contradict each other, and encoding this logic beyond simple rules required careful prompt design and structured AI outputs.

Finally, simulating a realistic reservation experience — without actually calling restaurants — required balancing realism with simplicity for a live demo.

Accomplishments that we’re proud of

  • Built an end-to-end AI-mediated group decision system
  • Successfully modeled real-world dining conflicts and resolved them transparently
  • Created a live, collaborative experience with real-time updates
  • Delivered a polished, demo-ready product that feels immediately useful

Watching a group reach consensus in under five minutes was a huge win.

What we learned

This hackathon provided deep, hands-on experience building AI-assisted social workflows:

  • Designing prompts that produce structured, explainable outputs
  • Orchestrating AI as a decision facilitator, not a decision maker
  • Deploying scalable, cloud-native systems on Google Cloud Run
  • Balancing realism with hackathon constraints

Most importantly, we learned that AI is most powerful when it helps people collaborate better.

What’s next

Social Dining has plenty of room to grow:

  • Real restaurant reservations and calendar integration
  • Voice-based group planning (“Hey AI, where should we eat tonight?”)
  • Smarter weighting of preferences based on group dynamics
  • Post-meal feedback loops to improve future recommendations
  • Expansion beyond dining into travel, movies, and events

Our long-term goal is to turn everyday group decisions into effortless, AI-guided experiences.

Built With

  • Next.js 14
  • React
  • Tailwind CSS
  • FastAPI
  • Python
  • Supabase (PostgreSQL)
  • Yelp AI API
  • Google Cloud Run
  • Docker

Built With

Share this project:

Updates