Google is expanding Gemini in Chrome from a subscriber-only feature into a broader part of everyday browsing for U.S. desktop users. The update brings the assistant to all Mac and Windows desktop users in the U.S. who have their language set to English, and it sets up Chrome for a larger shift toward AI-assisted search, browsing, security, and task completion.
The announcement covers both features users can access now and capabilities Google says are coming later. Together, they point to a browser that is meant to explain pages, compare tabs, recall past sessions, connect with Google apps, help with complex searches, and eventually carry out routine online tasks while leaving the final confirmation to the user.
Gemini in Chrome is moving beyond paid subscribers
Gemini in Chrome had previously been limited to Google AI Pro and Google AI Ultra subscribers. Google is now rolling it out to all Mac and Windows desktop users in the U.S. who use English as their language setting.
Users can open Gemini from the icon in the top-right corner of the Chrome window. From there, they can ask it to explain or clarify complex information on the web page they are reading.
The source example is simple but useful: a person viewing a banana bread recipe could ask Gemini to make the recipe gluten free. In practice, that shows how Google wants the assistant to work inside the current page rather than as a separate destination.
Gemini can also work across multiple tabs. Google says this allows users to summarize and compare information from different websites without manually switching back and forth through each page.
That could apply to trip planning, where flight, hotel, and vacation pages are open at the same time. It could also apply to shopping, such as comparing different mattress models across multiple tabs.
Chrome is preparing for more agentic browsing
Google also described upcoming agentic capabilities for Chrome. The company says Gemini will be able to handle tedious online tasks, including booking a haircut or ordering weekly groceries.
The important boundary is that Gemini will not complete the purchase on its own. Google says the assistant will navigate to the site, add items to the cart, and then let the user take the final action by checking out with their payment option.
That distinction matters because it places Gemini in the role of an assistant that can move through a website and prepare an action, while the user remains responsible for the final step. It also shows how Chrome is being positioned as more than a place to view pages: Google wants it to become a layer for coordinating work across websites.
Google says these new agentic capabilities will be available in Chrome in the coming months. The source also notes that OpenAI launched Operator, an AI agent that performs tasks autonomously, earlier this year.
Memory and Google app integrations are part of the plan
Gemini will soon be able to retrieve web pages a user has visited before. Google says this is meant to make it easier to return to past browsing sessions without digging through browser history.
The examples Google gave are conversational searches over past activity. A user could ask, “On which site did I see the walnut desk last week?” or “What was that blog I read on back-to-school shopping?”
Google is also launching deeper integration between Gemini in Chrome and other Google apps, including Calendar, YouTube, and Maps. The company says users will be able to schedule meetings, see location details, and do more without leaving the page they are on.
One example involves YouTube. If a user is trying to find a specific spot in a video, they can ask Gemini to take them there.
These examples show the same general direction as the tab and history features: Gemini is being designed to work across the things a person already has open, has recently viewed, or is using through Google’s own apps.
AI Mode is coming to the address bar
Google is also bringing AI Mode, its advanced search feature, directly into the Chrome address bar. The change is meant to make complex, follow-up-based searches available from the same place users already type web addresses and search terms.
Instead of entering a short query such as “best mattress,” a user could type a more detailed request directly into the address bar: “I’m a side sleeper with occasional lower back pain. Make me a table comparing the different mattress types”. From there, the user could continue with follow-up questions such as, “How long do memory foam mattresses typically last?”
Google says this update will roll out later this month in English in the U.S. It will expand to more countries and languages in the future.
The address bar will also gain the ability to ask questions about the page currently open. Chrome can suggest relevant questions based on the page context, giving users a starting point for a search. Google says users will receive a helpful AI Overview and the option to continue with follow-up questions in AI Mode.
Security updates target scams and exposed passwords
Google is also applying AI to browser security. The company says Chrome will soon use its Gemini Nano model to detect and protect against scams, including fake virus alerts and fraudulent giveaways.
Google notes that these scams often impersonate trusted brands and use generative AI to create convincing phishing attempts. In that context, the browser is being asked to recognize threats that may themselves be made more persuasive by AI tools.
Password protection is also getting an AI-assisted update. Google says it is using AI to help users fix compromised passwords with a single click on supported sites, including Coursera, Spotify, Duolingo, H&M, and more.
If Chrome warns that a password was exposed in a data breach, the user can allow Chrome to create and save a new one. That turns a warning into a more direct recovery flow, at least on supported sites.
Across these updates, Chrome is becoming a more active product. It is not only displaying pages, but also interpreting them, organizing them, searching from them, remembering them, and helping users respond to security problems inside the browser.