LxDIG provides AI coding assistants with a powerful, persistent code intelligence layer, transforming repositories into dynamic, queryable knowledge graphs. Moving beyond the limitations of static RAG, LxDIG offers live, incrementally-updated insights, empowering agents with structural code understanding, enduring memory, and safe multi-agent coordination. It's designed to solve critical challenges faced by AI coding agents, including context loss, architectural blindness, and conflicts in parallel workflows, ensuring they can understand, plan, implement, verify, and remember across sessions.
Key Features
01Multi-agent coordination with claim/release protocols and conflict detection for parallel workflows.
02Impact-scoped test selection and change analysis to run only affected tests.
031 GitHub stars
04Persistent agent memory to recall observations, decisions, edits, and learnings across sessions.
05Hybrid retrieval (Graph + Vector + BM25) fused with Reciprocal Rank Fusion (RRF) for accurate context.
06Code graph intelligence with natural-language and Cypher queries for structural understanding.
Use Cases
01Enable safe and efficient multi-agent workflows for engineering teams, preventing file conflicts and enhancing coordination.
02Empower individual developers with AI agents that understand complex code structures and remember past decisions.
03Accelerate CI/CD pipelines with impact-scoped test selection and automated architecture compliance checks.
