mcp s

This MCP repository is a curated collection of production-ready MCP servers that wrap open-source MCP implementations. Each server exposes a Model Context Protocol interface that lets AI agents or clients call tool-like capabilities (e.g., search, calculation, automation) through a consistent MCP API. You can run individual servers from this monorepo and connect them to your MCP client or agent workflow just like the examples shown in the README. The Brave Search MCP Server offers web search capabilities powered by Brave Search, while the Calculator MCP Server provides mathematical computations and evaluations. You can expand your setup by running more servers from the monorepo and registering them with your MCP client for unified access to multiple tools.

To use a server, start it as described in the installation steps, then configure your MCP client to point at the server by name. For example, Brave Search can be invoked via an npx-based command, and Calculator via a Python module invocation. The monorepo supports multiple transport options and authentication schemes, so you can tailor access control and response behavior (including standby mode for faster responses) to your needs.

Prerequisites:

• Node.js 18+ and npm (for TypeScript/Node-based servers)
• Python 3.8+ and pip (for Python-based servers)
• Docker (optional, for containerized deployment)

Installation steps (per server):

Brave Search MCP Server (TypeScript/Node):

1. Navigate to the server directory:
   • cd brave-search-mcp-server
2. Install dependencies:
   • npm install
3. Start the server:
   • npm start

Calculator MCP Server (Python):

1. Navigate to the server directory:
   • cd calculator-MCP-server
2. Install dependencies:
   • pip install -r requirements.txt
3. Run the server:
   • python -m src.main

Optional Docker deployment (illustrative):

• Build and run containers for individual servers as defined by their Dockerfiles
• Example (per server): docker build -t brave-search-mcp-server ./brave-search-mcp-server docker run -d --name brave-search-mcp-server -p 8080:8080 brave-search-mcp-server

Tips:

• You can mix and match servers in a single MCP client configuration by listing multiple mcpServers entries.
• Use the provided cwd in the client config if you run servers from subdirectories.

Environment variables and configuration options can vary by server. Common considerations:

• Ensure correct Node.js or Python versions as required by each server.
• If using Docker, expose appropriate ports and configure environment variables for authentication or rate limits.
• For Python servers, some projects may rely on virtual environments; consider using venv or pyproject.toml tooling if needed.
• The monorepo encourages running servers independently; each server typically has its own start script or entry point.
• When integrating with an MCP client, you can reuse the example configuration structure shown in the repository README, extending it with additional servers as you add them.