- Step 1: Install the MCP-Undetected-Chromedriver package via npm or Docker.
- Step 2: Start the MCP server with the command `mcp-server-undetected-chromedriver`.
- Step 3: Configure your client application to connect to the MCP server.
- Step 4: Use provided API functions like `browser_navigate`, `browser_screenshot`, `browser_click`, etc., to control the browser.
- Step 5: Perform desired automation tasks, then close the browser with `browser_close`.