Memory

Provides a persistent memory graph for Claude across chats, with entities, observations, and relations.

javascript

6

GitHub Stars

javascript

Language

6 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "yodakeisuke-mcp-memory-domain-knowledge": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "mcp/memory"
      ],
      "env": {
        "MEMORY_FILE_PATH": "/path/to/custom/memory.json"
      }
    }
  }
}

This MCP server provides a persistent knowledge graph memory for Claude interactions, letting it remember information about users across chats. It stores entities, their observations, and relations, so you can build a dynamic memory that improves personalization over time.

How to use

You can run the Memory MCP server locally and connect Claude Desktop to use a shared memory graph. Start the server via a container or through NPX, then interact with Claude as it reads from and writes to the memory graph. Use search, open, and update operations to store people, places, and facts, and let Claude retrieve relevant memories at the start of conversations.

How to install

Prerequisites: you need Docker or Node with NPX installed to run the Memory MCP server.

// Docker method
// Start memory server using Docker
docker run -i --rm mcp/memory

// NPX method
// Start memory server using NPX
npx -y @modelcontextprotocol/server-memory

// NPX with custom memory path (optional) MEMORY_FILE_PATH=/path/to/custom/memory.json npx -y @modelcontextprotocol/server-memory

## Additional setup details

If you want Claude Desktop to manage a custom memory file, you can supply MEMORY\_FILE\_PATH to the NPX-based server. The default memory storage is memory.json in the server directory.

## Available tools

### create\_entities

Create multiple new entities in the knowledge graph. Each entity has a name, entityType, and a list of observations. Existing names are ignored.

### create\_relations

Create multiple new relations between entities. Each relation has a from, to, and relationType in active voice. Duplicates are skipped.

### add\_observations

Add new observations to existing entities. Requires the target entity to exist. Returns added observations per entity.

### delete\_entities

Remove entities and their related relations. Performs cascading deletion of associated relations. Silently skips non-existent entities.

### delete\_observations

Remove specific observations from entities. Silently skips if the observation does not exist.

### delete\_relations

Remove specific relations from the graph. Silently skips if the relation does not exist.

### read\_graph

Read and return the entire knowledge graph, including all entities and relations.

### search\_nodes

Search nodes by keywords across names, types, subdomains, and observations. Supports case-insensitive, partial matches with OR semantics for multiple keywords.

### open\_nodes

Retrieve specific nodes by name along with their interconnecting relations. Non-existent nodes are skipped.