← Back to Packages

ContextMemory: MCP Server & 3D Visualizer 🧠

An ultra-lightweight, lightning-fast Model Context Protocol (MCP) Server and 3D Visualizer designed to act as a plug-and-play package. You can use this repository to instantly create your own custom MCP memory server for any project—simply by plugging in your own Neo4j Aura database credentials!

🌟 Features

  • Plug-and-Play MCP Server: Instantly give your AI coding assistants (Antigravity, OpenCode, Claude Code, Cursor, etc.) a persistent, shared memory via search_memory, add_memory, and delete_memory tools.
  • Lazy-Loaded PyTorch: Instant server boot times (less than 4s). AI models only load into RAM when a memory operation is explicitly requested.
  • Realtime 3D Visualizer: A sleek local web application that maps your project's memory as a glowing, interactive neural network.
  • Neo4j Aura Ready: Directly offloads graph storage and vector queries to Neo4j Cloud, meaning it works on any machine without local database installation.
  • 100% Secure: Credentials are kept entirely out of the code using .env files, ensuring your database passwords are never leaked to GitHub!

🚀 Quick Start (Local Setup)

  1. Clone the Repository

    git clone https://github.com/mmyahya29/MCP-Visualizer.git
cd MCP-Visualizer

  2. Configure your Database Credentials Rename the .env.example file to .env and insert your Neo4j Cloud credentials:

    NEO4J_URI=neo4j+s://...
NEO4J_USERNAME=neo4j
NEO4J_PASSWORD=your_password

  3. Launch the Visualizer!

    • Windows: Double-click run.bat
    • Mac/Linux: Run ./run.sh

    This automatically sets up your Python virtual environment, installs the requirements, and opens the 3D visualizer at http://localhost:8000.

🤖 Connecting to your AI Agents (MCP Setup)

To give your AI coding agents access to this brain, configure them to use this repository as a local MCP server.

For Antigravity: Add the following to your ~/.gemini/config/mcp_config.json:

"mcpServers": {
  "foster-memory": {
    "command": "/absolute/path/to/MCP-Visualizer/venv/Scripts/python.exe",
    "args": [
      "/absolute/path/to/MCP-Visualizer/server.py"
    ]
  }
}

(Mac/Linux users: change venv/Scripts/python.exe to venv/bin/python)

For OpenCode: Add the following to your ~/.config/opencode/opencode.json:

"mcp": {
  "foster-memory": {
    "type": "local",
    "command": [
      "/absolute/path/to/MCP-Visualizer/venv/Scripts/python.exe",
      "/absolute/path/to/MCP-Visualizer/server.py"
    ],
    "enabled": true
  }
}

🌐 Deploying to Vercel

If you want to host your 3D Visualizer permanently on the internet:

  1. Copy the contents of the VercelVisualizer folder into a new GitHub repository.
  2. Go to Vercel.com and deploy that specific repository.
  3. In your Vercel Dashboard, add your NEO4J_URI, NEO4J_USERNAME, and NEO4J_PASSWORD environment variables.

Built to make your AI tools smarter.

View on GitHub