Zanny's Persistent Memory Manager

A custom MCP (Model Capabilities Provider) server that functions as a persistent memory bank. This TypeScript-based server allows you to store, retrieve, and manage memories with flexible keyword detection and commands.

Features

• Store Memories: Save any text-based information for later retrieval
• Retrieve Memories: Search and retrieve stored memories by content or ID
• Delete Memories: Remove memories when they're no longer needed
• Smart Keyword Detection: Automatically detect relevant commands in natural language
• Unlimited Storage: No artificial limits on memory storage size
• Comprehensive Logging: Detailed logging for troubleshooting and monitoring
• MCP Compatibility: Full JSON-RPC implementation for Smithery.ai deployment

Installation

1. Ensure you have Node.js and npm installed on your system
2. Install project dependencies:

   npm install
   

3. Build the TypeScript project:

   npm run build
   

Usage

Starting the Server

Start the server with:

npm start

The server will begin listening on port 3000 by default.

API Endpoints

REST API (Legacy)

• GET /health - Check server health
• POST /api/memories - Store a new memory
• GET /api/memories - List or search memories
• GET /api/memories/:id - Retrieve a specific memory
• DELETE /api/memories/:id - Delete a memory
• POST /api/detect - Detect trigger keywords in text

JSON-RPC Endpoints (MCP Compatible)

• POST /tools/list - List available tools
• POST /tools/call - Call a specific tool

Memory Commands

The MCP server understands natural language commands related to memory management. Examples:

Storing Memories

remember: This is important information I want to store
store: The meeting is scheduled for March 15th at 2pm

Retrieving Memories

recall information about meetings
remember anything related to schedules

Deleting Memories

delete memory with id 5f4dcc3b5aa765d61d8327deb882cf99
forget id 5f4dcc3b5aa765d61d8327deb882cf99

Listing All Memories

list all memories
show all memories

Project Structure

├── dist/                # Compiled JavaScript output
├── logs/                # Log files
├── src/                 # TypeScript source code
│   ├── config.ts        # Server configuration
│   ├── index.ts         # Entry point
│   ├── logger.ts        # Logging configuration
│   ├── mcpServer.ts     # Main MCP server implementation
│   └── memoryManager.ts # Memory storage and retrieval
├── package.json         # Project dependencies
├── smithery.json        # Smithery.ai configuration
├── tsconfig.json        # TypeScript configuration
└── README.md            # This file

Configuration

Configuration is managed in src/config.ts. The main settings include:

• Server name
• Server port
• Memories directory
• Logger configuration
• Trigger keywords

Deploying to Smithery.ai

This MCP server is compatible with Smithery.ai deployments. To deploy:

1. Make sure your code is in a Git repository
2. Create a Smithery.ai account if you don't have one
3. Connect your repository to Smithery.ai
4. The included smithery.json file will guide the deployment process

Logging

Logs are stored in the logs/ directory. All logs are formatted as JSON to ensure compatibility with the MCP specification.

License

ISC