Gemini in Chrome can now see exactly what you’re looking at on screen

Google has added a "Select from screen" feature to Gemini in Chrome, letting users highlight specific text or images on a webpage and send them directly to the AI assistant. Rolling out with Chrome 149, the tool removes the need to manually describe what's on screen. Google also gave developers access to Gemini 3.5 Flash's computer-use capabilities, enabling AI agents to interact with browsers and apps for complex, multi-step tasks.