browser-use-mcp-server

MCP.Pizza Chef: kontext-security

Instead of pasting links and copying text back and forth, this hands the whole browsing job to your assistant: it opens pages, clicks through them, fills forms, and reports what it found. It runs on your own computer and works with Claude Desktop, Cursor, and Windsurf. You supply an OpenAI key, and installation involves Python tooling or Docker, so plan on a technical helper for setup. A built-in viewer lets you watch the browser as it works.

Web/Research

Use This MCP server To

Open a news site and bring back the top story Fill out a long web form without typing it myself Read a page that only loads after clicking around Compare prices across a few shopping sites in one go Watch the browser work live instead of trusting a summary

README

browser-use-mcp-server

An MCP server that enables AI agents to control web browsers using browser-use.

🌐 Want to Vibe Browse the Web? Open-source AI-powered web browser - Vibe Browser.

🔗 Managing multiple MCP servers? Simplify your development workflow with agent-browser

Prerequisites

uv - Fast Python package manager
Playwright - Browser automation
mcp-proxy - Required for stdio mode

# Install prerequisites
curl -LsSf https://astral.sh/uv/install.sh | sh
uv tool install mcp-proxy
uv tool update-shell

Environment

Create a .env file:

OPENAI_API_KEY=your-api-key
CHROME_PATH=optional/path/to/chrome
PATIENT=false  # Set to true if API calls should wait for task completion

Installation

# Install dependencies
uv sync
uv pip install playwright
uv run playwright install --with-deps --no-shell chromium

browser-use-mcp-server FAQ

Which apps does this work with?

Claude Desktop, Cursor, and Windsurf are the ones the project documents, and any other app that speaks MCP can connect too.

Do I need a paid key?

Yes. It uses an OpenAI key that you supply, so every browsing session bills to your own OpenAI account.

Can I use this to collect information from a site with no download button?

Yes. It drives a real browser, so it can reach anything you could reach by clicking through yourself.

How hard is the setup?

This sits at the technical end of the scale. You install Python tooling and a browser engine, or run the whole thing in Docker.

Can I see what it is doing?

Yes. The project ships a live viewer, so you can watch the browser window while a task runs.

Is it still being worked on?

Yes. The project now lives under the kontext-security account and its code changed as recently as May 2026.