Fetch With Images

🚀 MCP Fetch Tool

The Model Context Protocol server is used to fetch web content and process images. This enables Claude Desktop (or other MCP clients) to fetch web page content and handle images appropriately.

🚀 Quick Start (Users)

To use this tool with Claude Desktop, add the following to your Claude Desktop configuration file (~/Library/Application Support/Claude/claude_desktop_config.json):

{
  "tools": {
    "fetch": {
      "command": "npx",
      "args": ["-y", "@kazuph/mcp-fetch"]
    }
  }
}

This will automatically download and run the latest version of the tool as needed.

Required Settings

1. Enable Claude's accessibility features:
   • Open System Settings.
   • Go to Privacy & Security > Accessibility.
   • Click the "+" button.
   • Add Claude from the Applications folder.
   • Enable the toggle switch next to Claude.

This accessibility setting is required for automatic clipboard operations (Cmd+V).

✨ Features

• Web Content Extraction: Automatically extract and format web page content into Markdown.
• Article Title Extraction: Extract and display the article title.
• Image Processing: Optionally process images on web pages with optimization features (off by default; set enableFetchImages: true to enable).
• Pagination Support: Support pagination for text and images.
• JPEG Optimization: Automatically convert images to JPEG format and control quality.
• GIF Support: Extract the first frame of animated GIFs.

👨‍💻 Developer Information

The following section is for those who wish to develop or modify this tool.

📋 Prerequisites

• Node.js version 18 or higher.
• macOS system (due to dependencies on macOS-specific clipboard operations).

📦 Installation

Follow these steps to install:

npm install -g @kazuph/mcp-fetch

🔧 Image Processing Specifications

Image processing uses the Sharp library for optimal performance and quality. When multiple images are found, they are vertically stitched in order, taking size limitations into account.

⚙️ Configuration

Add the following configuration to ~/.config/mcp-fetch.json:

{
  "proxy": "http://your-proxy-server:port",
  "timeout": 30000,
  "maxConcurrency": 5
}

📄 API Documentation

Request Body Parameters

Basic Parameters

Advanced Parameters

Response Body Parameters

Basic Response

Image Response

💻 Usage Examples

Basic Usage

{
  "url": "https://example.com"
}

Enable Image Fetching

{
  "url": "https://example.com",
  "enableFetchImages": true,
  "imageMaxCount": 3
}

Paged Image Fetching

{
  "url": "https://example.com",
  "enableFetchImages": true,
  "imageStartIndex": 3,
  "imageMaxCount": 3
}

📌 Notes

• This tool is designed specifically for macOS due to its dependence on macOS-specific clipboard operations.
• Image processing uses the Sharp library to ensure high performance and quality.
• When multiple images are found, they are vertically stitched in order, considering size limitations.
• Animated GIFs are automatically processed to extract their first frame.

📝 Change Log

v1.1.3

• Modified default behavior: Do not fetch images by default (enableFetchImages: false).
• Removed the disableImages parameter and replaced it with the enableFetchImages parameter.

v1.1.0

• Added the article title extraction feature.
• Improved the response format to include the article title.
• Fixed type issues in the MCP response content.

v1.0.0

• Initial version released.
• Provided features such as web content fetching, image processing and optimization, and pagination support.