Memory

What it is

Knowledge Graph Memory Server is a basic implementation of persistent memory using a local knowledge graph, letting Claude remember information about the user across chats. It stores entities (name, type, observations), directed relations in active voice, and atomic observations. It exposes nine tools plus a knowledge-graph MCP resource, and is published on npm as @modelcontextprotocol/server-memory.

Who it's for

• Claude Desktop users who want Claude to remember user information across chats
• VS Code users configuring MCP servers
• Developers building chat personalization with a custom system prompt

Requirements

Setup

1. Claude Desktop via NPX

   Add this to your claude_desktop_config.json.

   jsonCopy

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

2. Claude Desktop via Docker

   Alternatively, add this to your claude_desktop_config.json to run the server in Docker with a named volume.

   jsonCopy

   {
     "mcpServers": {
       "memory": {
         "command": "docker",
         "args": ["run", "-i", "-v", "claude-memory:/app/dist", "--rm", "mcp/memory"]
       }
     }
   }

3. Custom storage file path

   Set MEMORY_FILE_PATH in the env block to change where the JSONL memory file is stored.

   jsonCopy

   {
     "mcpServers": {
       "memory": {
         "command": "npx",
         "args": [
           "-y",
           "@modelcontextprotocol/server-memory"
         ],
         "env": {
           "MEMORY_FILE_PATH": "/path/to/custom/memory.jsonl"
         }
       }
     }
   }

4. VS Code via NPX

   Add to your user mcp.json (Command Palette: MCP: Open User Configuration) or to .vscode/mcp.json in your workspace.

   jsonCopy

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

5. Build the Docker image

   Build the image locally from the repository.

   shCopy

   docker build -t mcp/memory -f src/memory/Dockerfile .

Examples

{
  "name": "John_Smith",
  "entityType": "person",
  "observations": ["Speaks fluent Spanish"]
}

{
  "from": "John_Smith",
  "to": "Anthropic",
  "relationType": "works_at"
}

{
  "entityName": "John_Smith",
  "observations": [
    "Speaks fluent Spanish",
    "Graduated in 2019",
    "Prefers morning meetings"
  ]
}

Follow these steps for each interaction:

1. User Identification:
   - You should assume that you are interacting with default_user
   - If you have not identified default_user, proactively try to do so.

2. Memory Retrieval:
   - Always begin your chat by saying only "Remembering..." and retrieve all relevant information from your knowledge graph
   - Always refer to your knowledge graph as your "memory"

Pros & cons

Images