Google rolls out Select from screen to let Gemini in Chrome analyze cropped page regions

What the new feature does

Google has begun rolling out a Gemini in Chrome capability called Select from screen, a desktop workflow that lets users hand the AI assistant a specific part of a web page instead of an entire page or a written description. According to dev.to, the feature targets situations where the full page is not the relevant context: users draw a bounding box around a passage of text, an image, or a region that mixes both, and Gemini then analyzes or acts on that selection in Chrome's side panel.

The report frames the value as context selection. Rather than explaining in words what is on screen, or copying material between tools, users point the assistant directly at the material they want considered — something that matters for teams that research products, review creative assets, compare suppliers, or work from web-based documents.

How it works

Google's official instructions, as described by dev.to, lay out a short process: the user opens the Gemini side panel in Chrome, picks Select from screen, and outlines the region to share. That content then forms the basis of the next exchange with the assistant.

One detail worth noting is that the selection is not tied to a single content type. Google says the chosen region can hold text, images, or a combination of the two, giving users a way to supply context from a page without making the whole page the prompt.

Rollout and access constraints

Select from screen belongs to the wider Gemini in Chrome initiative, which also covers tighter integrations with Google services including Calendar, Gmail, Maps, and YouTube. dev.to reports that Google has signaled more expansion for the feature family, from broader regional availability to added capabilities such as editing images found on the web.

Availability, though, is not uniform. The documentation the report draws on confirms the workflow but does not pin down a single worldwide release date, a universal price, or a guarantee that every Gemini in Chrome capability ships with every account tier. Google has indicated that some functions may hinge on rollout phase, region, or a Google AI Pro or Ultra subscription.

Managed environments add another layer. For Workspace domains and enterprise Chrome fleets, administrators may control whether Gemini in Chrome is available at all. dev.to suggests organizations verify whether the feature is enabled in Chrome's "AI innovations" settings, whether account, region, or subscription conditions apply, whether policy permits the tool, and what internal guidance should exist before staff feed sensitive page content to an AI assistant.

Limitations to keep in mind

Cropping a region is efficient, but a tight selection can strip away surrounding context that changes what the material means. And as with any AI output, Gemini's answers still warrant review before use in customer-facing, financial, legal, or operational decisions. Less friction in handing over context does not replace human oversight of the result.

For a practical evaluation, dev.to recommends a repeatable task with a clear expected outcome — selecting a product comparison area, for instance, and asking Gemini to identify the differences — then comparing the time saved against manually preparing the same prompt.

Why it matters

Much of the working day happens in the browser, and Select from screen moves AI assistance closer to the material people are actually looking at. Cutting out the describe-it-in-words step lowers the barrier to quick analysis of a chart, a listing, or a passage of text. For Google, it embeds Gemini more deeply into the browsing surface and differentiates it from standalone chatbots. For organizations, the rollout raises questions — availability, admin policy, and data handling — that are worth answering before the feature settles into routine workflows.