Neuromcp
Persist semantic memory for your coding agent with hybrid search, a knowledge graph, and consolidation so long projects stay coherent.
Overview
neuromcp is a MCP server for the Build phase that provides semantic agent memory with hybrid search, a knowledge graph, and consolidation over stdio.
What is this MCP server?
- Semantic memory for AI agents with hybrid search across embeddings and structured cues
- Knowledge graph relationships for linking entities, decisions, and code context
- Memory consolidation to merge and refine stored facts over long builds
- npm distribution with npx runtime hint (v0.4.2) and stdio MCP transport
- Fits solo builders who need durable project memory beyond chat context windows
- Server version 0.4.2
- npm package identifier neuromcp with npx runtime hint
- Transport: stdio
Community signal: 4 GitHub stars.
What problem does it solve?
Long solo builds lose critical decisions in chat history, forcing agents to relearn architecture and preferences every session.
Who is it for?
Builders running extended agent-assisted development who need graph-aware recall and hybrid search beyond the context window.
Skip if: One-off weekend prototypes with no need for persistent agent memory or graph maintenance.
What do I get? / Deliverables
After adding neuromcp, your agent can store, search, and relate project knowledge with consolidated semantic memory across sessions.
- Hybrid search MCP tools for agent recall
- Knowledge graph-linked memory entries
- Consolidation workflows for refined long-term context
Recommended MCP Servers
Journey fit
Agent long-term memory is configured during Build when you wire MCP tooling, then reused across Ship and Operate; canonical shelf is agent-tooling in Build. neuromcp is semantic memory infrastructure for agents—hybrid retrieval and graph consolidation—not a generic database admin tool.
How it compares
Semantic memory MCP with graph and consolidation, not a SQL database MCP or plaintext notes skill.
Common Questions / FAQ
Who is neuromcp for?
It is for solo and small-team developers who want durable, searchable semantic memory wired into their MCP coding agents.
When should I use neuromcp?
Use it when a project spans many sessions and you need hybrid retrieval plus structured relationships between stored facts.
How do I add neuromcp to my agent?
Configure a stdio MCP server using npx neuromcp (v0.4.2) in Claude Code, Cursor, or another MCP host, then grant agents tools to write and query memory per your retention policy.