Unified MCP Client Library
@pietrozullo
Basic information
About Unified MCP Client Library
mcp-use is the easiest way to interact with mcp servers with custom agents
Overview
What is Unified MCP Client Library?
Unified MCP Client Library (MCP-Use) is an open-source Python library that connects any LLM to any MCP server, enabling developers to build custom agents with tool access without using closed-source or application clients.
How to use Unified MCP Client Library?
Install via pip (pip install mcp-use) and the appropriate LangChain provider package for your chosen LLM (e.g., pip install langchain-openai). Configure MCP servers in a JSON file or dictionary, then create an MCPAgent with your LLM and an MCPClient to run queries.
Key features of Unified MCP Client Library
- Create your first agent in only six lines of code
- Works with any LangChain-supported LLM that supports tool calling
- Supports HTTP connections to MCP servers running on specific ports
- Agents can dynamically select the best server for a task
- Use multiple MCP servers simultaneously in a single agent
- Restrict potentially dangerous tools like file system or network access
Use cases of Unified MCP Client Library
- Web browsing and scraping using Playwright MCP server
- Search and book accommodations via Airbnb MCP server
- Create 3D models in Blender using Blender MCP server
- Combine multiple server tools (e.g., web search + file ops) in a single workflow
FAQ from Unified MCP Client Library
What is Unified MCP Client Library and how does it differ from closed-source clients?
It is an open-source Python library that lets any LLM use MCP tools, without relying on proprietary clients. Developers can build and customize agents with full control.
Which LLMs are supported?
Any LLM that supports tool calling and is integrated with LangChain, such as OpenAI, Anthropic, Groq, and Llama. Install the corresponding LangChain provider package.
Does it support HTTP-based MCP servers?
Yes. You can connect to MCP servers running on specific HTTP ports by providing a url field in the server configuration (e.g., "url": "http://localhost:8931/sse").
Can I restrict which tools the agent can use?
Yes. The library includes tool restriction features to limit access to dangerous capabilities like file system or network operations.
How do I manage multiple MCP servers?
Use the MCPClient with a configuration file listing all servers. The agent can access tools from all servers, and you can optionally enable dynamic server selection (use_server_manager=True) or specify a server name per task.
More Agent Frameworks MCP clients
MCP Client Chatbot
cgoingloveπ Open source MCP Client: A Multi-provider AI Chatbot Solution
Systemprompt Multimodal MCP Client
ejb503A Multi-modal MCP client for voice powered agentic workflows
MCP CLI client
adhikaspA simple CLI to run LLM prompt and implement MCP client.
π MCPOmni Connect - Universal Gateway to MCP Servers
Abiorh001MCPOmni Connect is an agent execution runtime that connects to multiple MCP servers via stdio, SSE, or streamable HTTP. It supports chat, autonomous agents, and planner-based orchestration treating each MCP server as a tool agent to enable dynamic multi-agent workflows across LLM
Novu - Connect your AI agents & products to customers
NovuNovu is the open-source communication layer for AI agents. An agent built on Novu connects to MCP servers like Linear, GitHub, and Notion for its tools, then talks with your users on Slack, Microsoft Teams, WhatsApp, Tel
Comments