MCP Catalogs
Homemcp-browser-agent screenshot

mcp-browser-agent

by imprvhub·38·Score 45

An MCP server providing Claude with autonomous browser automation capabilities including DOM manipulation, JavaScript execution, and API requests.

browser-automationweb-scrapingdeveloper-tools
12
Forks
2
Open issues
3 mo ago
Last commit
2d ago
Indexed

Overview

MCP Browser Agent is a comprehensive Model Context Protocol integration that extends Claude Desktop's capabilities with autonomous browser automation. Built on Playwright, it enables AI agents to interact with web content through various tools including navigation, screenshot capture, DOM manipulation, and JavaScript execution. The server maintains persistent browser state, captures console logs, and provides a full-featured API client for making HTTP requests. This allows Claude to perform complex web automation tasks with intelligent error recovery and multi-step instruction following.

Try asking AI

After installing, here are 5 things you can ask your AI assistant:

you:Automating web testing and QA processes through AI agents
you:Web scraping complex sites with dynamic content and JavaScript rendering
you:Automating repetitive browser workflows like form submissions and data extraction
you:Which browsers are supported?
you:How do I change the browser type?

When to choose this

Choose this when you need Claude to autonomously interact with web applications, perform data extraction, or automate browser-based workflows.

When NOT to choose this

Avoid if you need browser automation outside of Claude Desktop, or if you require headless browser operations without visual debugging capabilities.

Tools this server exposes

12 tools extracted from the README
  • browser_navigate

    Navigate to a URL

  • browser_screenshot

    Capture screenshot

  • browser_click

    Click element

  • browser_fill

    Fill form input

  • browser_select

    Select dropdown option

  • browser_hover

    Hover over element

  • browser_evaluate

    Execute JavaScript

  • api_get

    GET request

  • api_post

    POST request

  • api_put

    PUT request

  • api_patch

    PATCH request

  • api_delete

    DELETE request

Comparable tools

puppeteer-mcpplaywright-mcpbrowserless

Installation

Installation

  1. Clone the repository:
git clone https://github.com/imprvhub/mcp-browser-agent
cd mcp-browser-agent
  1. Install dependencies:
npm install
  1. Build the project:
npm run build
  1. Configure Claude Desktop:

Add the following to your claude_desktop_config.json:

{
  "mcpServers": {
    "browserAgent": {
      "command": "node",
      "args": ["ABSOLUTE_PATH_TO_DIRECTORY/mcp-browser-agent/dist/index.js", "--browser", "chrome"]
    }
  }
}

Replace ABSOLUTE_PATH_TO_DIRECTORY with the actual path to the project directory.

FAQ

Which browsers are supported?
The MCP Browser Agent supports Chrome (default), Firefox, Microsoft Edge, and WebKit (Safari engine). Browser drivers are automatically installed when first used.
How do I change the browser type?
You can specify the browser type through configuration file, command line argument, environment variable, or Claude Desktop configuration. See the Browser Selection section in the README for detailed instructions.

On Hacker News

Recent discussion from the developer community.

Compare mcp-browser-agent with

GitHub →

Last updated · Auto-generated from public README + GitHub signals.