What it does
The Memory server is one of the reference servers in the Model Context Protocol repository, maintained by the MCP steering group. It provides a knowledge graph-based persistent memory system that gives LLMs access to stored information through MCP. It is written in TypeScript.
It suits developers who want to study how an MCP server exposes data and tools to a model, or who want an example to base their own server on. The publisher describes the repository's servers as educational reference implementations that demonstrate MCP features and SDK usage, not as production-ready solutions.
The server is started with npx by an MCP client, which is configured with a command and arguments. It uses the stdio transport, so it opens no network port and waits for the client to send messages. On Windows, the npx command must be wrapped with cmd /c. The MCP Inspector can launch the server and let you browse and call its tools without a full client setup.
Limitations and requirements
The publisher describes this as a reference implementation intended for learning, not a production-ready solution. Developers must assess their own security requirements and add safeguards suited to their threat model and use case. It needs an MCP client to be configured to launch it, because it runs over stdio and opens no port.
Safety
Our team has not reviewed this resource. Only automated rules checked it; resources with high-risk signals are not published. These are signals, not a guarantee that installing it is safe. Try it in a sandbox project first.
- Runs code or changes an assistant's behaviour
Sources
- Cline MCP Marketplace first seen 13 Nov 2026 · last seen 18 Nov 2026 · its page on the source
92,041
As shown on cline.bot when we last read it