Knowledge Graph Generator
Turns PDF documents into an evolving knowledge graph through a containerized, idempotent pipeline.
Upload PDFs, and a worker extracts their content and grows a knowledge graph you can explore in the browser.
- Pipeline. Unstructured-IO parses PDFs, GraphRAG-SDK builds the graph, and FalkorDB stores it.
- Scaling. Kafka queues work asynchronously, MinIO stores the files, and Redis hashes keep every stage idempotent.
- Interface. A FastAPI backend and a Streamlit frontend, all orchestrated with Docker Compose.