AI Agent with MCP

What is AI Agent with MCP?

AI Agent with MCP is an implementation of an AI agent that uses the Model Context Protocol (MCP) to interact with external resources. It includes a functional MCP server with REST resources for users and messages, along with a client to consume those resources. The project is designed for developers who want to build and test MCP servers with NodeJS, particularly for integration with the Cursor editor.

How to use AI Agent with MCP?

Install the server via npx -y @smithery/cli@latest install @wonderwhy-er/desktop-commander --client cursor --key <your-key>, or manually add it to your mcp.json file (Mac/Linux or Windows). You can also clone the repository and run npm run setup. After installation, use the provided TypeScript client code to connect and list resources, or access resources directly via REST (e.g., curl http://3.238.149.189:8080/users).

Key features of AI Agent with MCP

• Provides api://users resource listing all registered users.
• Provides hello://world resource returning a simple greeting.
• Includes a functional MCP client built with the official SDK.
• Supports installation via Smithery CLI or manual Cursor config.
• Allows direct REST API access to resources.

Use cases of AI Agent with MCP

• Retrieving a list of users with ID, phone number, name, and thread ID.
• Getting a “Hello, World!” greeting message through an MCP resource.
• Learning how to create and consume MCP resources in NodeJS.
• Integrating MCP resources into the Cursor editor environment.
• Testing MCP server-client communication with practical examples.

FAQ from AI Agent with MCP

How do I install the MCP server?

You can install it using npx with Smithery CLI: npx -y @smithery/cli@latest install @wonderwhy-er/desktop-commander --client cursor --key <your-key>. Alternatively, add the server configuration manually to mcp.json (located at ~/.config/cursor.json on Linux), or clone the repository and run npm run setup.

What resources are available?

The server provides two resources: api://users (returns a JSON list of users with id, phoneNumber, name, and threadId) and hello://world (returns a plain text greeting “Hello, World! This is my first MCP resource.”).

How do I use the MCP client?

Import Client and StdioClientTransport from the SDK, create a client with name: "hello-mcp-client", connect using new StdioClientTransport({ command: "node", args: ["build/index.js"] }), then call client.listResources() and client.readResource({ uri: "api://users" }).

What are the system requirements?

Node.js (version compatible with AbortController), TypeScript, and the @modelcontextprotocol/sdk package are required.

How can I access resources via REST?

Run curl http://3.238.149.189:8080/users to get the list of users directly over HTTP. The hello resource is accessed via the MCP protocol only.