Browser Use MCP Server

About

Browser Use MCP lets AI assistants run browser automation through local stdio or the separately configured Browser Use Cloud service.

Key Features

• Control browser sessions from MCP-compatible AI tools.
• Use a local installation or delegate browser tasks to managed cloud infrastructure.

Use Cases

Navigate websites, collect page information and exercise browser workflows through an agent.

Limitations

Local runtime setup, model credentials and cloud billing depend on the selected deployment.

Sources

• Project documentation: https://docs.browser-use.com/open-source/customize/integrations/mcp-server
• Setup reference: https://docs.browser-use.com/cloud/guides/mcp-server

MCP Setup

Prerequisites

• Local: uv/uvx and Chrome or Chromium.

Access requirements

• Local: an OpenAI or Anthropic model API key.
• Cloud: Browser Use account and API key.

Client setup

Claude Code

Transport: stdio

claude mcp add browser-use -- uvx --from "browser-use[cli]" browser-use --mcp

Local deployment. Make the model key available to the server process.

Cursor

Transport: http

{
  "mcpServers": {
    "browser-use": {
      "url": "https://api.browser-use.com/v3/mcp",
      "headers": {
        "x-browser-use-api-key": "YOUR_API_KEY"
      }
    }
  }
}

Cloud deployment; uses Browser Use Cloud credentials.

Environment variables

• OPENAI_API_KEY or ANTHROPIC_API_KEY (local deployment)

Setup docs: https://docs.browser-use.com/open-source/customize/integrations/mcp-server