Google’s latest AI model uses a web browser like you do

Gemini 2.5 Computer Use uses visual understanding and reasoning capabilities to analyze a users request and carry out a task. Can be used for UI testing or navigating interfaces made for people who dont have an API or other direct connection available.