Google has introduced a handy "Select from screen" tool for the Gemini assistant in Chrome. Users simply click the tool to highlight the current tab and freely select any text or image, and the selected content is automatically added to the Gemini prompt. This feature is rolling out gradually as part of Chrome 149—if you don't see it right away, restarting the browser should do the trick. It greatly simplifies bringing on-screen information into AI conversations, making queries more intuitive and efficient.
Meanwhile, Google announced that the Gemini 3.5 Flash model now comes with built-in computer use capabilities. Developers can call the model via the Gemini API to interact with interfaces and analyze app content just like a human would. For example, 3.5 Flash can automatically open the Gemini app and return a categorized list of functions. This progress marks a further step in AI agent capabilities, opening up new possibilities for automated testing, intelligent assistants, and more.
