Scrappey MCP Server

A Model Context Protocol (MCP) server for interacting with Scrappey.com's web automation and scraping capabilities. Try it out directly at smithery.ai/server/@pim97/mcp-server-scrappey.

Overview

This MCP server provides a bridge between AI models and Scrappey's web automation platform, allowing you to:

Create and manage browser sessions
Send HTTP requests through Scrappey's infrastructure
Execute browser actions (clicking, typing, scrolling, etc.)
Handle various anti-bot protections automatically

Setup

Get your Scrappey API key from Scrappey.com
Set up your environment variable:

SCRAPPEY_API_KEY=your_api_key_here

Available Tools

1. Create Session (`scrappey_create_session`)

Creates a new browser session that persists cookies and other state.

{
  "proxy": "http://user:pass@ip:port"  // Optional: Custom proxy, leave empty for default
}

2. Destroy Session (`scrappey_destroy_session`)

Properly closes a browser session.

{
  "session": "session_id_here"  // Required: The session ID to destroy
}

3. Send Request (`scrappey_request`)

Send HTTP requests through the Scrappey infrastructure.

{
  "cmd": "request.get",  // Required: request.get, request.post, etc.
  "url": "https://example.com",  // Required: Target URL
  "session": "session_id_here",  // Required: Session ID to use
  "postData": "key=value",  // Optional: POST data
  "customHeaders": {  // Optional: Custom headers
    "User-Agent": "custom-agent"
  }
}

4. Browser Actions (`scrappey_browser_action`)

Execute browser automation actions.

{
  "session": "session_id_here",  // Required: Session ID to use
  "browserActions": [  // Required: Array of actions to perform
    {
      "type": "click",  // Action type: click, hover, type, scroll, wait
      "cssSelector": ".button",  // CSS selector for element
      "text": "Hello",  // Text to type (for type action)
      "wait": 1000  // Wait time in ms
    }
  ]
}

Typical Workflow

Create a session:

{
  "name": "scrappey_create_session"
}

Use the returned session ID for subsequent requests:

{
  "name": "scrappey_request",
  "cmd": "request.get",
  "url": "https://example.com",
  "session": "returned_session_id"
}

Perform browser actions if needed:

{
  "name": "scrappey_browser_action",
  "session": "returned_session_id",
  "browserActions": [
    {
      "type": "click",
      "cssSelector": "#login-button"
    },
    {
      "type": "type",
      "cssSelector": "#username",
      "text": "myuser"
    }
  ]
}

Clean up by destroying the session when done:

{
  "name": "scrappey_destroy_session",
  "session": "returned_session_id"
}

Features

Session persistence for maintaining state
Automatic anti-bot protection handling
Support for custom proxies
Browser automation capabilities
HTTP request methods (GET, POST, PUT, DELETE, PATCH)
Custom headers and cookies support

Best Practices

Always destroy sessions when you're done with them
Reuse sessions when making multiple requests to the same site
Use appropriate wait times between actions for more human-like behavior
Check if a session exists before using it

Error Handling

The server will return error messages with details when something goes wrong. Common errors include:

Invalid session IDs
Network timeouts
Invalid selectors for browser actions
Anti-bot protection failures

Resources

License

MIT License

Integrations