Inspiration
We wanted to create a smarter, more personal way for people to discover culture. Instead of generic recommendations, we imagined a tool that understands individual taste using AI.
What it does
CultureMatch AI takes a user's favorite movie, book, or artist and uses the Qloo API and GPT-4 to generate intelligent, personalized cultural recommendations, optionally based on location.
How we built it
- Backend: Python + FastAPI
- Frontend: HTML + Tailwind CSS
- APIs: Qloo Taste API + OpenAI GPT-4
- Environment: .env for API key management
- Hosted locally, structured for cloud deployment
Challenges we ran into
- Managing API keys securely
- Prompt tuning to keep GPT responses focused
- Rate limiting on external APIs
- Designing a clean and intuitive UI under time pressure
Accomplishments that we're proud of
- Fully functional AI-powered recommendation system
- Smooth integration between two APIs
- Sleek UI with minimal design effort
- Personalized results that feel human-curated
What we learned
- API integration and key management best practices
- Effective prompt engineering with GPT-4
- Building lightweight, fast UIs with Tailwind
- Structuring multi-component AI projects
What's next for CultureMatch AI
- Spotify and YouTube Music taste syncing
- Location-aware event discovery
- Image input support (e.g., posters, covers)
- Browser extension or chatbot versions
Log in or sign up for Devpost to join the conversation.