# Browserless MCP for AI Agents MCP

> Browserless lets your AI agent automate web tasks by controlling headless Chrome directly. You can take high-quality screenshots, generate polished PDFs from any URL, extract fully rendered content from JavaScript sites, and run targeted scraping requests—all through natural conversation.

## Overview
- **Category:** developer-tools
- **Price:** Free
- **Tags:** headless-chrome, pdf-generation, screenshot-capture, web-scraping, automation, cloud-browser

## Description

When you connect this MCP to your preferred AI client, it gives your agent the ability to interact with the live web browser. You stop treating websites as static text blocks; instead, you treat them like actionable data sources. Need a visual record? Just ask for a screenshot of a specific page or an element. Want reports? The system converts entire webpages into clean, professional PDFs. And if you need structured information, it can run custom scraping scripts to pull out exactly what you're looking for—like all product titles or key dates. Since Vinkius hosts this MCP in the catalog, your agent gets immediate access to these web automation tools alongside everything else you use.

## Tools

### check_system_health
Checks if the Browserless service is currently running and available for use.

### generate_pdf
Creates a downloadable PDF document using all the content from a specified URL.

### get_account_config
Retrieves your specific account settings and configuration details for verification.

### get_page_content
Fetches the fully rendered HTML source code from a given URL, including JavaScript-loaded elements.

### get_usage_stats
Retrieves metrics about your account usage history and current quotas.

### list_active_sessions
Lists all browser sessions that are currently running through the service.

### run_scrape
Executes a custom scraping script based on element selectors to extract specific data points.

### take_screenshot
Captures an image of any URL using headless Chrome, supporting full-page views.

## Prompt Examples

**Prompt:** 
```
Can you grab a full-page screenshot of that article at example.com/blog/post-1?
```

**Response:** 
```
**Screenshot Status:** ✅ Complete

**URL Captured:** `https://example.com/blog/post-1`
**Capture Type:** Full Page
**Dimensions:** 1920 x 3540 pixels

[Image Placeholder: High-res screenshot of the article]

The image is attached, showing the header through to the comments section. Let me know if you need a cropped version.
```

**Prompt:** 
```
I need product titles and prices from this shop page.
```

**Response:** 
```
I ran the scraping script on that URL and pulled out the data based on your request. Here are the top 5 results:

| Product Title | Price |
| :--- | :--- |
| Leather Jacket, Deluxe | $499.00 |
| Denim Jeans, Slim Fit | $89.99 |
| Cotton T-shirt, Basic | $19.99 |
| Running Shoes, Ultra X | $125.00 |
| Wool Scarf, Winter Blend | $55.00 |

Would you like me to export this list into a CSV file?
```

**Prompt:** 
```
Turn the entire content of this research paper into a PDF.
```

**Response:** 
```
**PDF Generation Status:** ⚙️ Processing...

Please wait while I compile all assets and formatting from `https://example.com/research-paper`.

*Success!*
Your document is ready for download: [Download Link].
The resulting file, 'Research Paper.pdf', maintains the original formatting including tables and embedded images.
```

## Capabilities

### Capture Web Visuals
Take detailed screenshots of any URL, including options for full-page captures.

### Generate Documents from URLs
Convert the content of an entire webpage into a polished PDF document.

### Extract Rendered Web Content
Retrieve the complete HTML code, even from complex websites that rely heavily on JavaScript to load their data.

### Targeted Data Scraping
Run specific scraping requests by defining element selectors and extracting only the necessary pieces of data.

### Monitor Browser Activity
List currently active browser sessions or check general service health status.

## Use Cases

### Documenting a Client Website
A Content Manager needs to provide a client with a report on their latest marketing blog post. They ask the agent to run `generate_pdf` on the URL, getting a single, clean, branded document ready for immediate review.

### Debugging Front-End Code
A Web Developer suspects a JavaScript component isn't loading data correctly. They ask the agent to run `get_page_content` on the page, allowing them to examine the full rendered HTML and find the missing element.

### Gathering Competitive Data
A market analyst needs a list of product titles from 5 competitor shop pages. They instruct the agent to use `run_scrape`, specifying selectors for 'product title' across all five URLs, compiling one structured list.

### QA Visual Regression Testing
A QA Tester runs a workflow that uses `take_screenshot` on key landing pages before and after a code deployment. They can compare the resulting images to ensure no visual elements shifted unexpectedly.

## Benefits

- Generate polished reports instantly. Instead of manually printing or saving a webpage, use the `generate_pdf` tool to create professional PDF documents directly from your agent.
- Bypass JavaScript roadblocks. The `get_page_content` tool retrieves the full HTML source code, letting you access data from modern, dynamic websites that usually fail with basic scrapers.
- Automate visual verification. Use `take_screenshot` to grab high-quality images of specific URLs or entire pages for documentation and QA checks without touching a physical browser.
- Isolate specific data points. If you only need the names and prices from an e-commerce page, the `run_scrape` tool lets you target those exact elements using selectors.
- Maintain oversight. The `list_active_sessions` tool lets your agent check if background automation tasks are still running or have failed.

## How It Works

The bottom line is that your AI client runs complex browser commands—like taking screenshots or scraping data—without you having to open a single tab yourself.

1. Subscribe to this MCP and enter your personal Browserless.io API Token.
2. Your AI client connects to the Vinkius catalog, recognizing this web automation capability.
3. You issue a simple command through natural language (e.g., 'Take a screenshot of X URL'), and the agent executes the necessary headless Chrome tasks.

## Frequently Asked Questions

**How does Browserless MCP handle websites that load content with JavaScript?**
It retrieves the fully rendered HTML source code, meaning it doesn't just grab what appears initially. It processes the site as if a user had opened it in Chrome, allowing you to access dynamic data and interactive elements.

**Can I use Browserless MCP to make PDF reports from URLs?**
Yes, you can use the generate_pdf tool. Simply provide the URL, and the agent returns a polished, downloadable PDF document that preserves the original page's formatting.

**Is Browserless MCP only for taking screenshots?**
No. It’s much more than just visual captures. You can use it to scrape structured data, get full HTML content, and create documents all through natural language conversation.

**What if I need a list of specific items from a webpage?**
You can run custom scraping scripts. By telling the agent exactly what element selectors to look for (like product names or dates), it pulls out only that information and gives you a clean list.

**Are there limits when using Browserless MCP?**
You can check your usage limits by running the get_usage_stats tool. This provides clear metrics on your current account status so you know how much capacity you have for automation runs.