Chrome Gemini Adds Screen Select Feature, 3.5 Flash Now Controls Computers

Google is rolling out a practical new feature for the Gemini assistant in Chrome: “Select from screen.” This tool lets users quickly add any text or image from the current tab to the prompt box. Simply click “Select from screen” in the menu, highlight the current tab, choose your target area, and the selected content automatically fills into Gemini’s query box. The feature is being gradually pushed with Chrome 149—restart your browser if you don’t see it immediately.

At the same time, Google announced that the Gemini 3.5 Flash model now includes built-in “computer use” capabilities. Developers can call this via the Gemini API, allowing the model to perform actions like analysis and classification in a virtual environment. In the official demo, 3.5 Flash used this ability to “analyze the Gemini app and return a feature category list.” This marks a significant step from passive answering to active interface manipulation, opening new possibilities for automated testing, data collection, and more.

From screen selection to computer control, Google is gradually upgrading Gemini from a “conversational assistant” to an “environment-aware assistant.” The former improves daily user convenience, while the latter gives developers a more human-like automation approach. As these tools mature, AI will integrate more tightly with the desktop ecosystem.