mcpconn 0.1.2

Simplified MCP client wrapper for efficient server interactions

mcpconn: The Missing Connector for AI

mcpconn is a Python library that provides a simple and efficient way to connect your applications to AI models using the Model Context Protocol (MCP). It acts as a wrapper around the mcp library, offering a streamlined client interface for seamless integration with various AI providers and transport protocols.

Table of Contents

• mcpconn: The Missing Connector for AI
  • ✨ Features
  • 🚀 Getting Started
    • Installation
    • Quick Start
  • 📚 Documentation
  • 🗺️ Roadmap
  • 🤝 Contributing
  • 📄 License
  • ⚠️ Disclaimer
  • Code of Conduct
  • 🛡️ Security
  • 🌟 Showcase
  • 💬 Support

✨ Features

• Simplified Client Interface: A high-level MCPClient for easy interaction with MCP servers.
• Multi-provider Support: Out-of-the-box support for Anthropic and OpenAI models. Note: OpenAI only supports remote MCP endpoints (not local/stdio/localhost). See: https://platform.openai.com/docs/guides/tools-remote-mcp
• Flexible Transports: Connect to servers using STDIO, SSE, or Streamable HTTP. OpenAI only supports remote MCP endpoints.
• Built-in Guardrails: Protect your application with content filtering, PII masking, and injection detection.
• Conversation Management: Easily manage conversation history, context, and persistence.
• Asynchronous by Design: Built with asyncio for high-performance, non-blocking I/O.
• Extensible: Easily add new LLM providers, transports, or guardrails.

🚀 Getting Started

Installation

pip install mcpconn

Quick Start

Here's a simple example of how to use mcpconn to connect to an MCP server and interact with an AI model:

import asyncio
from mcpconn import MCPClient

async def main():
    # Connect to a local server using STDIO (Anthropic only)
    client = MCPClient(llm_provider="anthropic")
    await client.connect("python examples/simple_server/weather_stdio.py")

    # ---
    # OpenAI usage example (remote MCP only):
    # client = MCPClient(llm_provider="openai")
    # await client.connect("https://mcp.deepwiki.com/mcp", transport="streamable_http")
    # ---

    # Note: OpenAI does NOT support local/stdio/localhost servers. See: https://platform.openai.com/docs/guides/tools-remote-mcp

    # Start a conversation
    conversation_id = client.start_conversation()
    print(f"Started conversation: {conversation_id}")

    # Send a message and get a response
    response = await client.query("Hello, world!")
    print(f"AI: {response}")

    # Disconnect from the server
    await client.disconnect()

if __name__ == "__main__":
    asyncio.run(main())

📚 Documentation

For full details on all features and the complete API reference, please visit our documentation site.

The documentation is automatically generated from the main branch and includes:

• A full Getting Started guide.
• In-depth tutorials and examples.
• The complete API Reference.

🗺️ Roadmap

• Add support for more LLM providers.
• Implement a more comprehensive test suite.
• Add more examples and tutorials.
• Improve documentation and type hinting.

🤝 Contributing

Contributions are welcome! If you'd like to contribute to mcpconn, please follow these steps:

1. Fork the repository.
2. Create a new branch for your feature or bug fix.
3. Make your changes and add tests.
4. Ensure that the tests pass.
5. Submit a pull request with a clear description of your changes.

📄 License

mcpconn is licensed under the MIT License.

⚠️ Disclaimer

This project is under active development and may undergo significant changes.

Code of Conduct

We are committed to providing a welcoming and inclusive environment for everyone. Please read and follow our Code of Conduct.

🛡️ Security

If you discover a security vulnerability, please report it to us by emailing [e-mail]. We will address all reports promptly.

🌟 Showcase

Have you built something cool with mcpconn? Written an article or created a video? We'd love to see it! Please open a pull request to add your project to this list.

💬 Support

If you have questions or need help, please open an issue in the issue tracker.