browser-use-mcp

🌐 MCP server wrapping browser-use — let AI browse the web autonomously.

What it does

Exposes browser-use's AI web agent as MCP tools, allowing AI assistants to:

• Browse websites and extract information

• Fill forms and interact with pages

• Navigate complex web workflows

• Execute multi-step web tasks

Related MCP server: Cloudflare Playwright MCP

Installation

# Clone the repo
git clone https://github.com/ellllapie/browser-use-mcp.git
cd browser-use-mcp

# Install dependencies
pip install -r requirements.txt

# Copy environment template
cp .env.example .env
# Edit .env and add your API key

Environment Variables

Option 1: Use with relay service (e.g., 灵眸)

OPENAI_API_KEY=your_relay_api_key
OPENAI_BASE_URL=https://your-relay-service.com/v1

Option 2: Use official OpenAI API

OPENAI_API_KEY=your_openai_api_key
# OPENAI_BASE_URL not needed

Option 3: Use Browser Use Cloud

BROWSER_USE_API_KEY=your_browser_use_api_key

Option 4: Use other providers

ANTHROPIC_API_KEY=your_anthropic_api_key
# or
GOOGLE_API_KEY=your_google_api_key

Usage

As MCP Server (for Kelivo, Claude Desktop, etc.)

Add to your MCP client config:

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

Available Tools

• browse_web: Execute a web browsing task and return results

Deployment

Deploy to Railway, Render, or any Python hosting platform. Set environment variables in your hosting dashboard.

License

MIT