# Puppeteer MCP — Headless Chrome Server for AI Agents

> MCP server wrapping Google Puppeteer for headless Chrome automation. Navigate pages, screenshot, generate PDFs, scrape SPAs, and test web apps through AI agent commands. 2,000+ stars.

## Install

Merge the JSON below into your `.mcp.json`:

## Quick Use

Add to your `.mcp.json`:

```json
{
  "mcpServers": {
    "puppeteer": {
      "command": "npx",
      "args": ["-y", "@anthropic/mcp-puppeteer"]
    }
  }
}
```

Restart Claude Code. Ask: "Take a screenshot of https://tokrepo.com and describe the layout."

---

## Intro

Puppeteer MCP is a Model Context Protocol server that wraps Google Puppeteer to give AI agents full headless Chrome browser control with 2,000+ GitHub stars. Navigate pages, take screenshots, generate PDFs, fill forms, scrape single-page applications, and run automated tests — all through natural language. Unlike Playwright MCP which supports multiple browsers, Puppeteer MCP focuses on Chrome/Chromium with lighter setup. Best for developers who need AI-assisted browser tasks without installing Playwright. Works with: Claude Code, Cursor, any MCP client. Setup time: under 1 minute.

---

## Available Tools

### `puppeteer_navigate`
```
"Open https://news.ycombinator.com"
"Navigate to our staging environment at https://staging.example.com"
```

### `puppeteer_screenshot`
```
"Take a screenshot of the current page"
"Screenshot only the header section"
"Take a full-page screenshot including below-the-fold content"
```

### `puppeteer_click`
```
"Click the Login button"
"Click the third item in the navigation menu"
```

### `puppeteer_fill`
```
"Fill the search box with 'AI coding tools'"
"Enter test@example.com in the email field"
```

### `puppeteer_evaluate`
```
"Run document.querySelectorAll('a').length to count all links on the page"
"Get the page title using document.title"
```

### `puppeteer_pdf`
```
"Generate a PDF of this page"
"Create a PDF report of the dashboard"
```

## Use Cases

### Visual Testing
```
"Navigate to our app, take screenshots of the login, dashboard, and settings pages"
-> Agent captures visual snapshots for design review
```

### SPA Scraping
```
"Open this React app, wait for data to load, extract the product list"
-> Puppeteer renders JS, agent reads the DOM
```

### PDF Generation
```
"Open the invoice page for order #123 and generate a PDF"
-> Agent navigates, renders, exports PDF
```

### Key Stats
- 2,000+ GitHub stars
- Headless Chrome/Chromium
- Screenshots, PDFs, scraping
- JavaScript execution
- Form filling and interaction

### FAQ

**Q: What is Puppeteer MCP?**
A: An MCP server that gives AI agents headless Chrome control via Puppeteer for screenshots, scraping, testing, and PDF generation.

**Q: Puppeteer MCP vs Playwright MCP?**
A: Puppeteer is Chrome-only and lighter. Playwright supports Chrome, Firefox, and WebKit with more features. Choose Puppeteer for simplicity.

**Q: Is Puppeteer MCP free?**
A: Yes, open-source. Puppeteer is free and maintained by Google.

---

## Source & Thanks

> Built on [Puppeteer](https://github.com/puppeteer/puppeteer) by Google. Licensed under Apache 2.0.
>
> [puppeteer](https://github.com/puppeteer/puppeteer) — stars 90,000+

---

<!-- ZH -->

## 快速使用

添加到 `.mcp.json`:

```json
{
  "mcpServers": {
    "puppeteer": {
      "command": "npx",
      "args": ["-y", "@anthropic/mcp-puppeteer"]
    }
  }
}
```

---

## 简介

Puppeteer MCP 是一个基于 Google Puppeteer 的 MCP 服务器，GitHub 2,000+ stars。AI Agent 可控制无头 Chrome 浏览器：导航、截图、PDF 生成、表单填写和 SPA 抓取。比 Playwright MCP 更轻量。

---

## 来源与感谢

> Built on [Puppeteer](https://github.com/puppeteer/puppeteer) by Google. Licensed under Apache 2.0.

---
Source: https://tokrepo.com/en/workflows/efa97531-0468-4c2b-ab24-aff9b4090ac3
Author: MCP Hub