Fetch MCP Server (TypeScript)

A Model Context Protocol server that provides web content fetching capabilities with Server-Sent Events (SSE) support, written in TypeScript. This server enables LLMs to retrieve and process content from web pages, converting HTML to markdown for easier consumption.

Features

• Written in TypeScript for better type safety and developer experience
• Uses Cheerio for HTML processing and Turndown for markdown conversion
• Server-Sent Events (SSE) support for real-time content streaming
• Fully compatible with the Model Context Protocol
• Configurable user agent
• Automatic removal of script, style, noscript, iframe, and img tags for cleaner content

Available Endpoints

SSE Endpoint

• GET /sse - Establishes an SSE connection for streaming content
• POST /messages - Handles message processing for the SSE connection

The server runs on port 3030 by default.

Installation

Option 1: One-Click Installation via Glutamate App (Recommended)

The easiest way to install and set up the Fetch MCP Server is through the Glutamate app:

1. Download and install Glutamate
2. Open Glutamate and navigate to the Extensions section
3. Find Fetch MCP Server and click "Install"
4. The app will automatically configure everything for you and runs locally

Option 2: Manual Setup

If you prefer to set up the Fetch MCP Server manually, you have two options:

Global Installation

npm install -g @glutamateapp/fetchts

Once installed globally, you can run the server directly:

fetch

Using npx

You can run the server without installation using npx:

npx @glutamateapp/fetchts

From Source

1. Clone the repository
2. Install dependencies:

npm install

3. Build the project:

npm run build

4. Start the server:

npm start

For development with hot reload:

npm run dev

Usage

The server implements the Model Context Protocol's fetch tool with the following parameters:

• url (string, required): URL to fetch
• maxLength (number, optional): Maximum number of characters to return (default: 5000)
• startIndex (number, optional): Start content from this character index (default: 0)
• raw (boolean, optional): Get raw content without markdown conversion (default: false)

Example Response Format

{
  "content": [
    {
      "type": "text",
      "text": "Processed content here..."
    }
  ],
  "metadata": {
    "truncated": true,
    "totalLength": 10000
  }
}

Development

Prerequisites

• Node.js (v14 or higher)
• npm or yarn

Testing

npm test

Contributing

Contributions are welcome! Please feel free to submit pull requests.

License

This project is licensed under the MIT License - see the LICENSE file for details.