home / mcp / fastmcp demo typescript mcp server

FastMCP Demo TypeScript MCP Server

Provides a local MCP server exposing tools, resources, and prompts for testing and learning MCP.

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "gogouravr-fast-mcp": {
      "command": "node",
      "args": [
        "/path/to/fast-mcp/dist/index.js"
      ]
    }
  }
}

You can run a lightweight MCP server that exposes tools, resources, and prompts so an AI partner can securely call functions, read data, and use templates. This server runs locally via standard input/output and is designed to be easy to test with common MCP clients.

How to use

You interact with the MCP server using a compatible MCP client. Start the server locally and point your client to the server as configured by the client’s MCP settings. The server provides tools you can call, resources you can read, and prompts you can reference to guide AI conversations. In practice, you will list available tools, call a tool to perform a calculation or fetch data, read a resource, and request a predefined prompt for a guided interaction.

How to install

Prerequisites: Ensure you have Node.js and npm installed on your machine.

1. Install dependencies.

2. Build the project.

3. Run the server.

4. Optionally run in development mode for auto-reload.

Configuration and usage notes

{
  "mcpServers": {
    "fast_mcp_demo": {
      "command": "node",
      "args": ["/path/to/fast-mcp/dist/index.js"]
    }
  }
}

Additional content

You can test the server with a compatible MCP client by adding a configuration entry that points to the server. For example, a client can be configured to start the local MCP server using the node runtime and the built server entry. You can run the server in standard I/O mode, which uses your terminal for input and output.

Notes for testing with MCP clients

To verify basic functionality, list available tools, read resources, and fetch prompts using your MCP client. Start the server, then use the client’s interface to request tool metadata, invoke a tool like a calculator, read a demo resource, or obtain a greeting prompt. This helps confirm end-to-end interactions work as expected.