home / mcp / web browser mcp server

Web Browser MCP Server

An advanced web browsing server for the Model Context Protocol (MCP) powered by Playwright, enabling headless browser interactions through a flexible, secure API.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "random-robbie-mcp-web-browser": {
      "command": "python",
      "args": [
        "/path/to/your/server.py"
      ]
    }
  }
}

You run an advanced MCP server that lets you control a headless web browser via a secure, flexible API. This enables automated web interactions, full-page extraction, multi-tab browsing, and scripted actions to build powerful browser-based automation and data collection workflows within your MCP client apps.

How to use

You connect to the web browser MCP server from your MCP client and issue actions to navigate, interact with pages, extract content, and gather page information. You can perform common web automation tasks such as navigating to a URL, filling forms, clicking elements, taking screenshots, and retrieving links. The server supports multiple tabs, waits for navigation, executes JavaScript on pages, and returns detailed page information for your downstream logic.

Typical usage patterns include opening a page, performing form input, and extracting content or screenshots. You can also collect all links on a page or filter them by a pattern, then switch between tabs or close them as your workflow requires.

How to install

Prerequisites you need before starting:

Python 3.10+

MCP SDK

Playwright

Install the MCP package and Playwright dependencies, then install browser binaries.

pip install mcp playwright
playwright install

Claude Desktop configuration

Configure the MCP server in the Claude Desktop configuration file to point to your local server process.

{
  "mcpServers": {
    "web_browser": {
      "command": "python",
      "args": [
        "/path/to/your/server.py"
      ]
    }
  }
}

Available tools

browse_to

Navigate to a URL and establish a page context for subsequent actions like extraction, interaction, and scripting.

extract_text_content

Extract visible text from the current page or a specific element using an optional CSS selector.

click_element

Click a page element specified by a CSS selector to trigger events, forms, or navigation.

get_page_screenshots

Capture screenshots of the full page or a specific element for visual verification or reporting.

get_page_links

Retrieve all links from the current page, with an optional pattern filter to narrow results.

input_text

Type text into a form field identified by a CSS selector.

create_new_tab

Open a new browser tab, optionally navigating to a specified URL.

switch_tab

Switch the active tab to another one by its ID.

list_tabs

List all currently open tabs and their identifiers.

close_tab

Close a specific tab or the current tab if no ID is provided.

refresh_page

Refresh the current page to re-fetch content.

get_page_info

Obtain detailed metadata about the current page, including URL, title, and loading state.

scroll_page

Scroll the page in a specified direction by a given amount, such as page, half, or pixel values.

wait_for_navigation

Wait for in-page navigation to complete within a timeout period.

execute_javascript

Run custom JavaScript in the context of the current page and return the result.