Memory Graph MCP Server

Official
AI & Reasoning
stdio
Featured

Maintain persistent memory across AI agent sessions using a knowledge-graph data structure. Store entities, relationships, and observations that persist between conversations. Essential for building AI agents with long-term context.

Generator PresetSimulate in PlaygroundRaw .md

Quick Install

$ npx -y @modelcontextprotocol/server-memory

In-Depth Setup Guide

Complete setup walkthrough & security best practices

Read Guide

Configuration by Client

{
  "mcpServers": {
    "memory": {
      "command": "npx",
      "args": [
        "-y",
        "@modelcontextprotocol/server-memory"
      ]
    }
  }
}

Need a different client? Use the Config Generator for Windsurf, Cline, Antigravity, and more.

Available Tools (6)

create_entitiescreate_relationsadd_observationssearch_nodesopen_nodesdelete_entities

Memory Graph FAQ

What is the Memory Graph MCP server?

Maintain persistent memory across AI agent sessions using a knowledge-graph data structure. Store entities, relationships, and observations that persist between conversations. Essential for building AI agents with long-term context.

How do I install the Memory Graph MCP server?

Run: npx -y @modelcontextprotocol/server-memory

Which AI clients support the Memory Graph MCP server?

The Memory Graph MCP server works with Claude Desktop, Cursor, Codex CLI, Antigravity, Windsurf, and any MCP-compatible client using stdio transport.

What tools does the Memory Graph MCP server provide?

The server provides 6 tools: create_entities, create_relations, add_observations, search_nodes, open_nodes, delete_entities.