Puppeteer MCP
Official Puppeteer browser automation server — let AI control a real browser for scraping, screenshots, and UI testing.
// Add to your client
{
"mcpServers": {
"puppeteer": {
"command": "npx",
"args": ["-y", "@modelcontextprotocol/server-puppeteer"]
}
}
}Paste into your client's MCP configuration file.
// Try it now
"Take a screenshot of https://example.com and save it to ./screenshot.png"
- ! Chromium has to be installed — the first run may download a few hundred MB
- ! Running in headful mode opens real browser windows on your desktop
- ! Some sites block headless browsers — you may need to tweak user agents
API key or simple config
// When to use
- You need to take screenshots of web pages
- You want to generate PDFs from HTML or live pages
- You're doing local scraping that requires JavaScript execution
- You want browser automation without cloud costs
// When NOT to use
- You need cloud-hosted browsers — use browserbase-mcp
- You only need static HTML — fetch-mcp is simpler
- You need cross-browser testing — use playwright-mcp (supports more engines)
// Usage Scenarios
Page Screenshots
Capture a full-page or viewport screenshot of a URL.
Example prompt
"Take a full-page screenshot of https://example.com/dashboard and save it to ./screenshots"
PDF Generation
Convert a web page or HTML to PDF.
Example prompt
"Generate a PDF of https://example.com/report/2026-q1 and save it as q1-report.pdf"
Scraping with JS
Extract data from pages that require JavaScript to render.
Example prompt
"Go to this infinite-scroll product page, scroll to load 50 items, and extract their names and prices"
// About
Plain English
Some websites only show their content after JavaScript runs — think dashboards, web apps, anything with infinite scroll or login walls. A simple URL fetcher gets back an empty shell. To see the real page, you need an actual web browser. This server gives your AI an actual web browser. It's a copy of Chrome, controlled in the background by a tool called Puppeteer. Tell Claude 'take a screenshot of this dashboard,' 'save this report as a PDF,' 'log into this admin page and pull the list of paying customers,' or 'scroll the page until 50 products are loaded and grab their names and prices' — and the AI drives a real browser to do it. It's powerful and runs entirely on your own machine, so there's no third-party service in the loop. The trade-offs: the first install downloads a couple hundred megabytes of Chromium, scraping logic can break when sites redesign, and on long sessions the browser quietly eats CPU and memory. For one-off automation, screenshots, or testing your own web app, it's exactly the right tool.
The official Puppeteer MCP server provides browser automation capabilities to AI assistants. Navigate URLs, take screenshots, interact with web elements, fill forms, click buttons, and extract page content — using a real Chromium browser.
One of the most powerful MCP servers for web scraping, UI testing, and automated browser workflows.
// Use Cases
- Automate web scraping without writing code
- Take and analyze screenshots of web pages
- Fill and submit forms through AI interaction
- Automated UI testing of web applications
// Works With
// Also Consider
// Related Servers
Brave Search MCP
Official Brave Search integration — give AI real-time web search access via the Brave Search API for up-to-date information.
Fetch MCP
Official web fetch server — let AI retrieve content from any URL, including HTML to Markdown conversion for clean LLM consumption.
Git MCP
Official Git MCP server — read repository history, diffs, blame, and status through AI without leaving your conversation.