Browser agent
An agent that navigates and acts on websites through a controlled browser.
Browser agents use a driver (CDP, Playwright, an extension) to load pages, read the DOM or accessibility tree, and click, type and submit. They are the practical form of computer use for web tasks — research, form filling, purchasing, scraping. WebArena (812 tasks) and Mind2Web are the standard benchmarks. Key failure modes are dynamic layouts, authentication and anti-bot systems.