01Token-Budgeted Context Assembly: Get optimal subgraphs to fit LLM token limits.
02Knowledge Freshness Tracking: Utilizes lifecycle metadata (active, superseded, draft) for AI agents to assess currency.
03Full-Text Search with Graph Context: SQLite FTS5 search optionally includes graph neighbors in results.
04Graph Traversal: Perform BFS walks with directional control (outgoing, incoming, both).
05Live Sync: Chokidar file watcher keeps graph and search indexes updated in real-time.
061 GitHub stars