Project KG establishes a centralized knowledge graph for AI-assisted knowledge work, ingesting and structuring data from work trackers, git repositories, and markdown files. It constructs a rich, searchable network of decisions, patterns, and discoveries, exposing this intelligence to AI agents via an MCP server. This enables capabilities like sophisticated semantic search (combining full-text and vector similarity), insightful graph traversal for understanding context, and automated knowledge capture, empowering AI tools like Claude Code to operate with a deeper, more comprehensive understanding of your projects.
Key Features
01Local embedding model for offline operation without API keys
02Connectors to sync external data sources (e.g., WCP) into the graph
03Graph traversal to explore connected decisions and context
041 GitHub stars
05Semantic search across all knowledge (FTS + vector similarity)
06MCP server for direct integration with AI agents like Claude Code
Use Cases
01Automating the capture and retrieval of lessons learned, decisions, and patterns from development workflows.
02Integrating work tracking data and project artifacts for a unified, context-rich view accessible by AI.
03Empowering AI agents with comprehensive project knowledge for enhanced productivity and decision-making.
