Mcp-Use FAQs

Question 1

How do I get started with Mcp-Use?

Accepted Answer

First, install Mcp-Use and LangChain.js with your chosen LLM provider using npm, yarn, or pnpm. Then, configure your MCP servers and instantiate an MCPAgent with your LLM and MCPClient. Refer to the Quick Start section in the documentation for a detailed code example.

Question 2

What kind of tools can I connect to my agent using Mcp-Use?

Accepted Answer

You can connect a wide range of tools, including web browsing, file operations, 3D modeling, and more, depending on the MCP servers you configure. Mcp-Use provides the framework for accessing these tools via LangChain.js.

Question 3

What is Mcp-Use?

Accepted Answer

Mcp-Use is a TypeScript library that allows you to connect LangChain.js-compatible Large Language Models (LLMs) to MCP (Managed Code Platform) servers. This enables you to build flexible AI agents with dynamic tool access and multi-server support.

Question 4

What are the key features of Mcp-Use?

Accepted Answer

Key features include easy agent creation with TypeScript, compatibility with any LangChain.js-supported LLM that supports tool calling, direct SSE/HTTP connection to MCP servers, dynamic server selection, multi-server support, and the ability to restrict unsafe tool access.

Question 5

Can I restrict certain tools from being used by my agent?

Accepted Answer

Yes, Mcp-Use allows you to restrict unsafe tools like filesystem or network access by specifying a list of `disallowedTools` when instantiating the MCPAgent. This enhances the security of your AI agents.

Mcp-Use

About

Key Features

Use Cases