Browser Use

This automates browser interactions through a persistent daemon that keeps Chrome running between commands, giving you ~50ms response times instead of the usual multi-second browser startup. You get element indices from `browser-use state`, then click, type, or extract data using those indices. It handles the annoying parts well: connects to your existing Chrome profile to preserve logins, runs headless by default, includes cloud browser fallback for zero-config setups. The command chaining works reliably since the browser stays alive. Genuinely useful for web scraping, form automation, or any task where you need to interact with actual web pages rather than just fetch HTML.