Google Chrome's Gemini can now see what you're highlighting

Highlighting a confusing chart on a website, snapping a screenshot, saving it, and dragging it into a chatbot sidebar is friction we have all learned to tolerate. Google is now trying to kill that workflow entirely.

In Chrome 149, Google started rolling out a new feature called "Select from screen." It essentially puts eyes directly inside your browser so Gemini can see whatever you are pointing at.

Just point and ask

Accessible via the "+" menu in the Chrome Gemini side panel, the tool works just like a native screenshot selector. Instead of copying text or explaining layout details, you highlight any image or chunk of text on your current tab, and Chrome immediately attaches it to your prompt.

Google is rolling out the feature across Chrome 149 right now, though you might need to restart your browser to see it show up. The goal here is obvious: turn Gemini from a standalone chat window into an in-context copilot that knows what you are doing without being explicitly told.

Google also paired this release with a massive developer update on the same day. Gemini 3.5 Flash now supports native computer use, allowing agents to see, reason, and execute actions across web browsers, desktop apps, and mobile environments for complex, multi-step automation jobs like software QA and back-office tasks.

Why it matters

We think this seemingly tiny UI tweak signals where consumer AI is actually heading: context over raw horsepower. Having the smartest model on earth does not help if getting context into the prompt feels like filing taxes. Making screen selection zero-friction turns Gemini into a natural reading companion for research, code reviews, and shopping.

Strategically, Google is defending its most valuable piece of real estate: the browser. By embedding native visual grounding into Chrome alongside agentic computer use via Gemini 3.5 Flash, it builds a moat that standalone web wrappers simply cannot match.

Still, giving an assistant persistent visual access inside your daily workflow is going to make some people nervous. Google will need to make sure the boundary between "what I highlighted" and "what else is on my private screen" remains crystal clear.