Spotlight

Inspiration

We were frustrated with endless scrolling through business listings and complex filters. We wanted to make finding local businesses as easy as asking a friend. With Yelp's AI API, we built a conversational assistant that understands what you're really looking for.

What it does

Spotlight lets you discover local businesses through natural conversations. Just chat with our AI - "Find a quiet coffee shop for working" - and get personalized recommendations from Yelp's database. You can compare businesses, make reservations, and save your favorites, all through simple conversation.

How we built it

  • Frontend: React + TypeScript + Vite
  • Backend: Node.js + Express
  • AI: Yelp AI API for intelligent recommendations
  • Database: MongoDB for user data
  • Auth: JWT tokens + bcrypt
  • Deployed: Render (separate frontend/backend)

Challenges we ran into

  1. API Rate Limits - Solved with smart caching (5-min TTL) that reduced calls by 80%
  2. Location Coverage - Built intelligent detection to handle Yelp's US/Canada focus
  3. Deployment Structure - Separated backend/frontend for Render's architecture
  4. Type Consistency - Centralized TypeScript definitions across services

Accomplishments that we're proud of

  • Natural chat interface that feels like texting a friend
  • 80% reduction in API calls through optimization
  • Full-stack production deployment on Render
  • Secure authentication with saved businesses
  • Smart location handling for accurate recommendations

What we learned

  • Yelp AI API integration and optimization
  • Real-time chat UI/UX patterns
  • API caching strategies
  • Cloud deployment best practices
  • Balancing AI intelligence with user control

What's next for Spotlight

  • Voice input for hands-free discovery
  • Social features to share recommendations
  • Interactive maps and neighborhood exploration
  • Multi-language support for global expansion
  • Business owner dashboard and analytics

Built With

Share this project:

Updates