CrabRAG: Why Automated Assistants Need Graph Memory, Not More Tokens — Stephen Chin, Neo4j
Graph-based memory outperforms flat markdown files for persistent AI agent recall
Neo4j DevRel lead Stephen Chin argues that AI agents suffer from memory amnesia because current tools (including Claude Code) rely on flat markdown files for context, which fail to capture relationships and persist state across sessions. CrabRAG proposes replacing token-stuffing approaches with graph memory as the retrieval backbone for autonomous agents. The talk addresses a real pain point in production agentic systems but is partly a vendor pitch; the solution architecture is not fully captured in this excerpt.