I Built a Local Knowledge Graph That Tells AI Agents "What Breaks If I Change This?"
Every time I start a new AI coding session, the same thing happens: I re-explain my architecture, my conventions, my service dependencies. The AI doesn't remember. It doesn't know that store.py is imported by 8 other modules, that Milenko owns the auth service, or that postgres is a single point of failure.
So I built Knowledge Master — a local knowledge graph that gives AI agents permanent memory of your codebase.
What makes it different from "just another RAG"
Most RAG tools embed your code into vectors and return "similar chunks." That's fancy grep. It doesn't understand relationships.






