Personal Neo4j Memory Server

Provides persistent graph memories and unified search for AI assistants using Neo4j with Cloudflare Access integration.

javascript

GitHub Stars

javascript

Language

5 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": {
    "jeelidev-personal-neo4j-memory-server": {
      "command": "npx",
      "args": [
        "-y",
        "https://github.com/jeelidev/personal-neo4j-memory-server.git"
      ],
      "env": {
        "NEO4J_URI": "bolt://localhost:7687",
        "NEO4J_PASSWORD": "your-password",
        "NEO4J_USERNAME": "neo4j"
      }
    }
  }
}

This MCP server provides AI assistants with persistent, graph-based memory using a Neo4j database, enhanced for secure remote access via Cloudflare. You can store memories as interconnected nodes, run powerful searches, and organize data by project while enjoying reliable remote tunneling and zero-trust security.

How to use

You interact with the Memory server through an MCP client. Start by connecting your client to the memory MCP server, then switch to the project database you want to work with. Use memory-related tools to store new memories, search for relevant memories with semantic or exact matches, and create relationships between memories to build a navigable knowledge graph. You can perform batch operations to create memories and relations in a single action, and adjust the level of detail returned by responses depending on your current needs (minimal lists vs full data with relations). Use the Cloudflare Access tunnel to securely reach the remote server from anywhere.

How to install

Prerequisites: you need Node.js and npm installed on your machine. If you plan to run the server locally, ensure you have Docker available only if you choose to deploy Neo4j DozerDB in a container.

Step 1. Install the MCP server via GitHub installation (recommended for now) and start using the provided MCP command.

# Start the MCP memory server using the GitHub installation link
npx -y https://github.com/jeelidev/personal-neo4j-memory-server.git

Step 2. If you are configuring Claude to use the MCP, add the memory server configuration to Claude's MCP setup as shown in the example below.

{
  "mcpServers": {
    "memory": {
      "command": "npx",
      "args": ["-y", "https://github.com/jeelidev/personal-neo4j-memory-server.git"],
      "env": {
        "NEO4J_URI": "bolt://localhost:7687",
        "NEO4J_USERNAME": "neo4j",
        "NEO4J_PASSWORD": "your-password"
      }
    }
  }
}

Additional setup steps

If you plan to run Neo4j remotely, you will configure Cloudflare Access to securely tunnel the Bolt protocol. Ensure your Cloudflare Access tunnel is running and your local server is reachable through the public hostname you configure.

Available tools

memory_store

Create memories with observations and immediate relations in a single operation.

memory_find

Perform unified search and retrieval combining semantic search, ID lookup, date filtering, and graph traversal.

memory_modify

Update, delete, and modify memories, observations, and relations.

database_switch

Switch the active project database to isolate contexts.

Built by

VeilStrat

AI signals for GTM teams