MCPRisingofficial publisher

Memory MCP (officiel)

MCP : Standardises the connection between an AI application and systems that expose tools or data · requires tout client MCP

Reference knowledge graph memory server, stored in a local file.

Sufficient for discovering persistent memory, too simple for serious use.

GitHub · 91,545 ★

Overall score

7.3 /10

Security: one caveat, see the card

Utility

6.5

Reliability

10.0

Security

One caveat

Ease

1 click · Comfortable

Security reflects where data goes, requested permissions and incidents on record at the verification date. It is not a guarantee of zero risk: check the deployment mode and the connections you enable.

What it does, concretely

Language models forget everything at the end of each conversation. Memory MCP addresses this by providing a persistent memory space structured as a knowledge graph, stored in a local file on your machine.

In practice, the server exposes tools that the model can call to create entities, establish relationships between them, and attach observations to them. This data is preserved between sessions in a local file, allowing the model to retrieve information from one conversation to the next.

This tool is aimed at developers who want to experiment with persistent memory in their own MCP projects, or at curious users who want to understand technically how this type of memory works.

It is not suitable for production use: the official repository explicitly states that these servers are reference implementations intended for educational purposes, not solutions ready for a serious environment. Security and robustness remain the responsibility of the developer who draws inspiration from them.

What it offers

·Creation of named entities in a knowledge graph that persists between sessions.
·Establishment of typed relationships between entities in the graph.
·Addition of textual observations associated with each entity.
·Local storage in a file on the user's machine, with no dependency on an external service.
·Immediate launch via npx without prior installation: npx -y @modelcontextprotocol/server-memory.
·Compatible with any MCP client, including Claude Desktop, Claude Code and Cursor.
·Open source code, usable as a learning base for building your own MCP server.

Our tips

On Windows, wrap the npx command with cmd /c in your MCP client's JSON configuration to avoid launch errors.
Use the mcpServers key in the Claude Desktop configuration file to declare the server and make it available automatically at each session.
Review the source code in the repository before drawing inspiration from it for a project: it is designed to illustrate SDK features, not to be copied as-is into production.
If you want to customise the storage file location, examine the server's launch options in the source code, as no graphical configuration interface is provided.

Watch out

The official repository explicitly warns that these servers are not production-ready solutions: security and safeguards must be implemented by the developer according to their own threat model.
Storage in a local file offers few guarantees in terms of reliability or confidentiality of the memorised data.

In practice

Price: Free
Commercial use: Commercial use: yes
Available on: claude-code, cursor
Host: tout client MCP
Licence: MIT

Security and data

Rating: Security: one caveat, see the card
Your data: Data stays local
Maintenance: Active
Last release: 25 October 2026

Links and repository checked on 25 October 2026 · added on 1 November 2024

How to install Memory MCP (officiel)

1. Make sure Node.js is installed on your machine (required for npx). 2. Open the configuration file of your MCP client (for example Claude Desktop). 3. Add the following entry in the mcpServers section: {"memory": {"command": "npx", "args": ["-y", "@modelcontextprotocol/server-memory"]}}. 4. On Windows, replace command with cmd and prepend /c and npx to the args array. 5. Restart your MCP client for the server to be recognised. 6. Security note: this server is a reference implementation intended for educational purposes; do not store sensitive data in it without assessing the risks.

npx -y @modelcontextprotocol/server-memory

In the same category

Whole category Feed (memory, context) →