🌐 MCP-Use is the open source way to connect any LLM to any MCP server and build custom agents that have tool access, without using closed source or application clients.

💬 Get started quickly - chat with your servers on our hosted version! Try mcp-use chat (beta) .

mcp-use is an open source library that enables developers to connect any LLM to any MCP server, allowing the creation of custom agents with tool access without relying on closed-source or application-specific clients.

Here's a quick example of how you can use mcp-use:

import asyncio
from langchain_openai import ChatOpenAI
from mcp-use import MCPAgent, MCPClient

async def main():
    client = MCPClient(config= {"mcpServers":{
    "playwright": {
        "command": "npx",
        "args": ["@playwright/mcp@latest"],
        "env": {"DISPLAY": ":1"}
    }}})
    # Create LLM
    llm = ChatOpenAI(model="gpt-4o", api_key=...)
    # Create agent with tools
    agent = MCPAgent(llm=llm, client=client, max_steps=30)
    # Run the query
    result = await agent.run("Find the best restaurant in San Francisco")

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

➡️ Create your own with our Builder

• 💻 Open Source: Connect any LLM to any MCP server without vendor lock-in.
• ⚙️ Flexible Configuration: Support for any MCP server through a simple configuration system.
• ⚙️ Easy Setup: Simple JSON-based configuration for MCP server integration.
• 🤖 Universal LLM Support: Compatible with any LangChain-supported LLM provider.
• 🔌 HTTP Connection: Connect to MCP servers running on specific HTTP ports for web-based integrations.
• 🔀 Dynamic Server Selection: Agents can dynamically choose the most appropriate MCP server for the task.

• 📥 Installation: Install mcp-use and set up your environment
• 📖 Configuration: Learn how to configure mcp-use with your MCP server

We are constantly working to improve mcp-use. Check out what we're planning and suggest new features!