READMEmain

Puppeteer MCP

A Multi-Commander Protocol for automating browser tasks with Puppeteer. This tool allows you to:

Automatically launch browsers (Chrome/Firefox)
Scrape web content from any URL
Execute custom JavaScript in the browser context
Take screenshots of webpages or specific elements

Installation

# Clone the repository
git clone https://github.com/yourusername/puppeteer-mcp.git
cd puppeteer-mcp

# Install dependencies
npm install

Usage

Launch a browser

# Launch Chrome in non-headless mode (visible)
node src/index.js launch

# Launch Chrome in headless mode
node src/index.js launch --headless

# Launch Firefox
node src/index.js launch --browser firefox

# Launch Chrome with a specific profile
node src/index.js launch --profile myprofile

Scrape web content

# Scrape an entire webpage
node src/index.js scrape https://example.com

# Scrape specific elements using a CSS selector
node src/index.js scrape https://example.com --selector "article.content"

# Save scraped content to a file
node src/index.js scrape https://example.com --output results.json

Take screenshots

# Take a screenshot of a webpage (viewport only)
node src/index.js screenshot https://example.com

# Take a screenshot of the full page (including scrollable content)
node src/index.js screenshot https://example.com --fullPage

# Take a screenshot of a specific element
node src/index.js screenshot https://example.com --selector "#header"

# Save screenshot to a specified file
node src/index.js screenshot https://example.com --output my-screenshot.png

Execute JavaScript in the browser

# Execute inline JavaScript
node src/index.js execute "return document.title;" --url https://example.com

# Execute JavaScript from a file
node src/index.js execute ./scripts/extract-data.js --url https://example.com

Close browser

# Close any open browser instances
node src/index.js close

Examples

Example 1: Scrape product information from a website

node src/index.js scrape https://example.com/products --selector ".product-card" --output products.json

Example 2: Take a screenshot of a webpage

# Take a full page screenshot and save it to a specific file
node src/index.js screenshot https://example.com --fullPage --output example-full.png

# Take a screenshot of just the hero section
node src/index.js screenshot https://example.com --selector "section.hero" --output hero.png

Example 3: Execute a custom data extraction script

Create a script file scripts/extract-data.js (already included in this repo):

// Extract all links from the page
const links = Array.from(document.querySelectorAll('a')).map(a => ({
  text: a.innerText.trim(),
  href: a.href,
  title: a.title
}));

// Return the extracted data
return {
  title: document.title,
  links: links.slice(0, 10) // Return only the first 10 links
};

Then run:

node src/index.js execute ./scripts/extract-data.js --url https://example.com

Features

Automatically finds installed browsers or downloads them if not found
Maintains browser state between commands
Supports both headless and non-headless modes
Works with both Chrome and Firefox
Can scrape specific elements or entire pages
Takes screenshots of full pages or specific elements
Allows execution of custom JavaScript code

License

MIT