MCP Manager

MCP Manager

A flexible server that enables communication between AI models and tools, supporting multiple MCP servers and compatible with Claude, MCP Dockmaster, and other MCP clients.

Category
Visit Server

README

MCP Manager

A flexible Model Context Protocol (MCP) manager server that enables communication between AI models and tools. Compatible with Claude, MCP Dockmaster, and other MCP clients.

Features

  • MCP Server Management: Run multiple MCP servers and manage them from a single interface
  • Worker Mode: Operate as a worker for other MCP clients
  • Auto-Discovery: Automatically find MCP servers on GitHub and from Smithery packages
  • Tool Registry: Browse and install tools from a registry of MCP tools
  • WebSocket Communication: JSON-RPC over WebSockets for real-time communication
  • Headless Mode: Run in headless mode for API-first operation

Getting Started

Prerequisites

  • Node.js (v18+)
  • NPM (v8+)

Installation

# Clone the repository
git clone <your-repo-url>
cd <your-repo-directory>

# Install dependencies
npm install

# Create a .env file from the example
cp .env.example .env

# Edit .env to configure your MCP Manager
nano .env

Running the Server

# Run in development mode
npm run dev

# Build and run in production mode
npm run build
npm run start

Configuration Options

Configure your MCP Manager by editing the .env file:

# Server Configuration
PORT=5000                  # Main server port
HOST=0.0.0.0               # Host to bind to
NODE_ENV=development       # Environment (development/production)
REUSE_PORT=true            # Whether to reuse the port

# MCP Port Range
MCP_PORT_RANGE_START=50050 # Start port for MCP servers
MCP_PORT_RANGE_END=50150   # End port for MCP servers

# MCP Compatibility Settings
MCP_WORKER_MODE=false      # Enable worker mode
MCP_DISCOVERY_ENABLED=true # Enable auto-discovery of MCP servers
MCP_AUTOSTART_SERVERS=true # Automatically start MCP servers

# Smithery API Configuration
SMITHERY_API_KEY=your_smithery_api_key_here # API key for Smithery packages

Using MCP Manager with Claude and Other MCP Clients

MCP Manager is designed to be compatible with various MCP clients, including:

Connecting to Claude Desktop

  1. Run MCP Manager on your local machine
  2. Open Claude Desktop
  3. Go to Settings > Advanced > MCP Configuration
  4. Set the MCP Server URL to http://localhost:5000/ws
  5. MCP Manager will appear as an available server in Claude

Connecting to MCP Dockmaster

  1. Run MCP Manager on your local machine
  2. Open MCP Dockmaster
  3. Click "Add Server"
  4. Enter the server information:
    • Name: MCP Manager
    • Address: localhost
    • Port: 5000
  5. Click "Connect"

Using Worker Mode

MCP Manager can operate as a worker for other MCP clients:

  1. Set MCP_WORKER_MODE=true in your .env file
  2. Run MCP Manager
  3. Connect to MCP Manager from your MCP client
  4. MCP Manager will automatically discover and register available tools

Headless/API-First Operation

MCP Manager supports headless operation for API-first clients:

GET /?api_key=YOUR_API_KEY&headless=true&auto_connect=true

Query parameters:

  • api_key: Your API key (if required)
  • headless: Enable headless mode
  • auto_connect: Automatically connect to available MCP servers
  • ws_url: Specify a custom WebSocket URL
  • server_id: Connect to a specific server ID

API Reference

MCP Manager implements the JSON-RPC protocol over WebSockets. The main API endpoint is:

ws://localhost:5000/ws

Available Methods

  • ping: Test server connectivity
  • getServers: Get all registered servers
  • getServer: Get server by ID
  • toggleWorker: Toggle worker status for a server
  • getTools: Get all tools
  • getToolsByServer: Get tools for a specific server
  • activateTool: Activate a tool
  • deactivateTool: Deactivate a tool
  • getStats: Get server statistics

License

MIT

Recommended Servers

playwright-mcp

playwright-mcp

A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.

Official
Featured
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.

Official
Featured
Local
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.

Official
Featured
Local
TypeScript
VeyraX MCP

VeyraX MCP

Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.

Official
Featured
Local
graphlit-mcp-server

graphlit-mcp-server

The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.

Official
Featured
TypeScript
Kagi MCP Server

Kagi MCP Server

An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.

Official
Featured
Python
E2B

E2B

Using MCP to run code via e2b.

Official
Featured
Neon Database

Neon Database

MCP server for interacting with Neon Management API and databases

Official
Featured
Exa Search

Exa Search

A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.

Official
Featured
Qdrant Server

Qdrant Server

This repository is an example of how to create a MCP server for Qdrant, a vector search engine.

Official
Featured