MCP Servers Collection

What is MCP Servers Collection?

MCP Servers Collection is a suite of Model Context Protocol (MCP) servers designed for Cline, Roo, and Claude Desktop AI assistants. It enables AI agents to interact with local and remote resources through a standardized protocol.

How to use MCP Servers Collection?

Clone the repository and start all servers with docker-compose up -d. For manual installation, install dependencies per server with pip install -r requirements.txt and run each with python main.py. Integrate with Claude Desktop by adding entries to claude_desktop_config.json as shown in the README.

Key features of MCP Servers Collection

• Three dedicated servers for Cline, Roo, and Claude Desktop
• Internal workflow orchestration and memory management
• Web scraping, browser automation, and screenshot capture
• Local file operations, shell commands, and desktop notifications
• Docker-based network isolation and token-based authentication
• Unified JSON-RPC endpoints with event-driven architecture

Use cases of MCP Servers Collection

• Automate web scraping and form filling with browser control
• Execute agent tasks and complex workflows for AI assistants
• Manage files, run shell commands, and send desktop notifications
• Integrate LLMs like Ollama for on‑device AI interactions

FAQ from MCP Servers Collection

What servers are included in this collection?

Three servers: Cline MCP Server (port 8080, workflow orchestration), Roo MCP Server (port 8081, web scraping and automation), and Claude Desktop MCP Server (port 8082, local file operations and LLM integration).

What are the runtime requirements?

Docker is recommended for containerized deployment. Manual installation requires Python and pip, plus Playwright for the Roo server and Ollama for the Claude Desktop server.

How are the servers secured?

Security features include network isolation via Docker, token-based authentication, input validation and sanitization, and resource limits with health checks.

What transport protocol do the servers use?

The servers communicate through unified /mcp endpoints using the JSON-RPC protocol with an event-driven architecture.

Can I run the servers manually without Docker?

Yes. Clone the repository, install requirements per server with pip install -r requirements.txt, and start each server individually by running python main.py in its directory.