Inspiration: Every city has thousands of events but no brain to match the right ones to the right person. What it does: Builds a knowledge graph of your city’s events, venues, and reviews, then generates a personalized evening plan tuned to your age and vibe. How we built it: Cognee (GraphRAG) for the knowledge graph, Qwen3-32B via free Groq for planning, FastAPI + Streamlit, custom ontology with 4 entity types and 15 relationships. Challenges: Making Cognee’s generic extraction understand event-specific concepts like “vibe,” crowd demographics, and walkable venue sequences. Accomplishments: A two-tier knowledge base (global + per-user) with 6 multi-hop query templates, real-time TriggerWare monitoring, and 103 passing tests in 48 hours. What we learned: A custom ontology turns a generic graph into a domain expert — the same Cognee instance goes from “here are some events” to “this jazz bar skews 30s after 10pm.” What’s next: Multi-city expansion, live source connectors, group planning mode, and a mobile app so your city’s Event Brain is always in your pocket.
Built With
- ai
Log in or sign up for Devpost to join the conversation.