What is Browser Use Mcp

Browser-use-mcp is a Model Context Protocol server designed for automating browser tasks using the Browser Use API. It provides tools for running automation tasks, monitoring their status, and managing ongoing tasks.

Use cases

Use cases include automating repetitive web tasks, monitoring web applications, scraping data from websites, and executing complex browser interactions without manual input.

How to use

To use browser-use-mcp, you need to install it via Smithery or pip, and run the server with your Browser Use API key. You can execute tasks using the provided tools like ‘run_task’, ‘get_task’, and ‘get_task_status’.

Key features

Key features include the ability to run browser automation tasks with specific instructions, retrieve task details by ID, and check the status of ongoing tasks.

Where to use

Browser-use-mcp can be used in various fields such as web scraping, automated testing, data collection, and any scenario where browser automation is required.

Browser Use MCP server

Overview

A Model Context Protocol server for automating browser tasks using Browser Use API. This server provides tools to run browser automation tasks, monitor task status, and manage running tasks.

Prerequisites

• A Browser Use API key

to get a Browser Use API key, go to Cloud Browser Use and sign up.

Installation

Installing via Smithery

To install browser-use-mcp for Claude Desktop automatically via Smithery:

npx -y @smithery/cli install @mhazarabad/browser-use-mcp --client claude

The package is not published to PyPI. You’ll need to clone this repository and run it directly from source.

git clone https://github.com/mhazarabad/browser-use-mcp.git
cd browser-use-mcp

Installing via pip

pip install cloud-browser-use-mcp-server

Running the Server

Using uvx (recommended)

you can run the server using python:

python -m cloud_browser_use_mcp_server --api-key YOUR_BROWSER_USE_API_KEY

Tools

1. run_task

   • Run a Browser Use automation task with instructions and wait for completion
   • Input:
     • instructions (string): Instructions for the browser automation task
     • structured_output (string, optional): JSON schema for structured output
     • parameters (object, optional): Additional parameters for the task
   • Returns: Information about the created task including final output if wait_for_completion is True

2. get_task

   • Get details of a Browser Use task by ID
   • Input:
     • task_id (string): ID of the task to retrieve
   • Returns: Complete task information including steps and output

3. get_task_status

   • Get the status of a Browser Use task
   • Input:
     • task_id (string): ID of the task to check
   • Returns: Current status of the task

4. stop_task

   • Stop a running Browser Use task
   • Input:
     • task_id (string): ID of the task to stop
   • Returns: Confirmation of task being stopped

5. pause_task

   • Pause a running Browser Use task
   • Input:
     • task_id (string): ID of the task to pause
   • Returns: Confirmation of task being paused

6. resume_task

   • Resume a paused Browser Use task
   • Input:
     • task_id (string): ID of the task to resume
   • Returns: Confirmation of task being resumed

7. list_tasks

   • List all Browser Use tasks
   • Returns: List of all tasks with their IDs and statuses

8. check_balance

   • Check your Browser Use account balance
   • Returns: Account balance information

Prompts

1. browser_use_task
   • Run a Browser Use automation task
   • Input:
     • instructions (string): Instructions for the browser automation task
     • structured_output (string, optional): JSON schema for structured output
   • Returns: Formatted task details as conversation context

Claude Desktop

Add this to your claude_desktop_config.json after installing it with pip install cloud-browser-use-mcp-server:

License

This MCP server is licensed under the MIT License. This means you are free to use, modify, and distribute the software, subject to the terms and conditions of the MIT License. For more details, please see the LICENSE file in the project repository.