Neuromcp is a local-first MCP (Model Context Protocol) server providing semantic memory for AI agents, featuring hybrid search (vector, full-text, and graph), knowledge graph consolidation, verbatim recall, and crash-resilient persistence. It exposes 40+ memory-related tools including advanced storage and retrieval, semantic deduplication, contradiction detection, cognitive priming, and automatic consolidation of memory and project knowledge. Neuromcp integrates with AI assistant tools and IDEs/editors (Claude, Cursor, Windsurf, Copilot, JetBrains, Zed), and can work with Ollama for enhanced LLM-based semantic search. It's valuable for developers building persistent, context-aware, and project-aware AI assistants, as well as anyone requiring robust local memory for AI workflows.
Visit Neuromcp's official website for product details and getting started.