Skip to content
AI Dev Toolkit.
Esc
  • AI Token CounterCount tokens for GPT, Claude, Gemini, DeepSeek, Qwen and more.Tool
  • LLM API Cost CalculatorEstimate per-request, daily and monthly API costs.Tool
  • AI Model ComparisonCompare prices, context windows and features across models.Tool
  • AI Model Pricing PagesSpecs, real costs and cheaper alternatives for popular models.Tool
  • Context Window CheckerSee whether your text fits each model’s context window.Tool
  • Subscription vs API CalculatorFind out whether a chat plan or the API is cheaper for you.Tool
  • GPU / VRAM CalculatorCheck how much VRAM a local model needs and which GPUs fit.Tool
  • Prompt Caching CalculatorEstimate savings from prompt caching.Tool

MCP server · Reference servers

Knowledge Graph Memory MCP server

Gives an assistant a persistent local knowledge graph of entities, relations and facts that it can add to and search.

Maintainer
Model Context Protocol project (official)
Licence
Apache-2.0 AND MIT
Runs
Locally (npx)
Sign-in
None
Last activity
2026-10-11

Overview

What it does

The Memory server is a Node.js reference server from the MCP project. It stores what an assistant learns as a small knowledge graph in a JSONL file on your machine. Entities have a name, a type and a list of observations, which are single facts, and relations link two entities in active voice, such as works_at.

Use it when you want an assistant to remember project details or preferences across chats without a cloud service. Tools create and delete entities, relations and observations, read the whole graph, search by name, type or observation text, and open nodes by name. The graph is also exposed as a resource that notifies subscribed clients when it changes. Set MEMORY_FILE_PATH to choose where the file lives. Several clients can share one file: writes take a lock file so they take turns, though the README warns this needs a local filesystem.

Tools

What the assistant can do

  • Create entities and relations, and add observations to entities
  • Delete entities, relations or individual observations
  • Read the whole graph or open chosen nodes by name
  • Search entities by name, type or observation text
  • Expose the graph as a resource with change notifications

Install

Install Knowledge Graph Memory MCP server in your app

You need:

  • Node.js with npx
  • A writable folder for the memory file (set MEMORY_FILE_PATH to choose it)

Run in your project. --scope local keeps the servers private to you in this project; use user for every project.

Terminal
claude mcp add --scope local memory -- npx -y @modelcontextprotocol/server-memory
Native Windows (PowerShell)
claude mcp add --scope local memory -- cmd /c npx -y '@modelcontextprotocol/server-memory'

Want it alongside other servers? Open Knowledge Graph Memory Server in the MCP config generator and add the rest; it writes one file per app. The snippets follow the setup in the server’s README, checked on 2026-10-14.

Before you install

  • Everything the assistant stores is written in plain text to a local JSONL file; by default it sits in the server’s own directory, so set MEMORY_FILE_PATH to a place you control.
  • The assistant can delete entities and observations, and deleting an entity also removes its relations.
  • Sharing one memory file between clients relies on exclusive file creation; the README says a network filesystem may not be safe.
  • A local MCP server runs with your user account’s permissions. Install it only from the source linked above. See how to choose a safe server.

Why it’s listed

Why we list it

A simple, local and inspectable way to give an assistant memory between sessions, kept by the MCP project.

Related

All MCP serversMCP config generatorGuide: MCP servers in Claude CodeGuide: what is MCP?