Google is introducing a new tool called "Select from screen" for the Gemini assistant inside the Chrome browser. This feature lets users quickly add any text or image from the current tab directly into a conversation prompt. Simply highlight the tab, choose "Select from screen," and draw a selection box around the content you want — Gemini will automatically fill the input field with the selected image or text. Rolling out gradually with Chrome 149, the feature can be activated by restarting the browser if it doesn't appear immediately. This update significantly simplifies visual query workflows, allowing users to let AI understand specific elements on a page without manually taking screenshots or copying and pasting.
At the same time, Google has announced that the Gemini 3.5 Flash model now comes with a built-in computer use tool. Developers can call this capability directly through the Gemini API, enabling the model to simulate human interactions with interfaces — such as browsing an app's UI and performing analysis tasks. In an official demo, 3.5 Flash used this tool to "analyze the Gemini app and return a categorized list of features." This means the model is no longer limited to understanding text; it can now "see" and "operate" graphical interfaces like a human, opening new possibilities for automated testing, smart assistants, and more. Gemini 3.5 Flash's computer use feature is available immediately in the Gemini API.
