Skrape MCP Server

Convert webpages into clean, LLM-ready Markdown using skrape.ai. An MCP server that seamlessly integrates web scraping with Claude Desktop and other MCP-compatible applications.

Key Features

• Clean Output: Removes ads, navigation, and irrelevant content
• JavaScript Support: Handles dynamic content rendering
• LLM-Optimized: Structured Markdown perfect for AI consumption
• Consistent Format: Uniform structure regardless of source

Features

Tools

• get_markdown - Convert any webpage to LLM-ready Markdown
  • Takes any input URL and optional parameters
  • Returns clean, structured Markdown optimized for LLM consumption
  • Supports JavaScript rendering for dynamic content
  • Optional JSON response format for advanced integrations

Installation

Installing via Smithery

To install Skrape MCP Server for Claude Desktop automatically via Smithery:

npx -y @smithery/cli install @skrapeai/skrape-mcp --client claude

Manual Installation

1. Get your API key from skrape.ai

2. Install dependencies:

npm install

1. Build the server:

npm run build

1. Add the server config to Claude Desktop:

On MacOS:

nano ~/Library/Application\ Support/Claude/claude_desktop_config.json

On Windows:

notepad %APPDATA%/Claude/claude_desktop_config.json

Add this configuration (replace paths and API key with your values):

{
  "mcpServers": {
    "skrape": {
      "command": "node",
      "args": ["path/to/skrape-mcp/build/index.js"],
      "env": {
        "SKRAPE_API_KEY": "your-key-here"
      }
    }
  }
}

Using with LLMs

Here's how to use the server with Claude or other LLM models:

1. First, ensure the server is properly configured in your LLM application
2. Then, you can ask the ALLMI to fetch and process any webpage:

Convert this webpage to markdown: https://example.com

Claude will use the MCP tool like this:

skrape
get_markdown

{
  "url": "https://example.com",
  "options": {
    "renderJs": true
  }
}

The resulting Markdown will be clean, structured, and ready for LLM processing.

Advanced Options

The get_markdown tool accepts these parameters:

• url (required): Any webpage URL to convert
• returnJson (optional): Set to true to get the full JSON response instead of just markdown
• options (optional): Additional scraping options
  • renderJs: Whether to render JavaScript before scraping (default: true)

Example with all options:

skrape
get_markdown

{
  "url": "https://example.com",
  "returnJson": true,
  "options": {
    "renderJs": false
  }
}

Development

For development with auto-rebuild:

npm run watch

Debugging

Since MCP servers communicate over stdio, debugging can be challenging. We recommend using the MCP Inspector:

npm run inspector

The Inspector will provide a URL to access debugging tools in your browser.