Gemini in Chrome on Android Explained: Who Gets It, What It Costs, and the Phone Requirement

Google opened Gemini in Chrome to all Android users in the United States on August 18, 2026. The browser assistant that summarises the page you are reading is now on your phone, it costs nothing, and paying subscribers get something considerably bigger: auto browse, which clicks through real websites and completes tasks for you.

The catch is that Android has the tightest rules of any platform Gemini in Chrome runs on. Desktop reaches 187 countries and territories and accepts 13-year-olds. Android is the United States only, in English only, for adults only, and it will not start on a phone that falls below a specific hardware line.

Key Takeaways

• It is free and it is live now. Google's post says Gemini in Chrome "is now available to all Android users in the U.S."
• United States only, English only. Google's availability page is blunt about it: "On Android devices, Gemini in Chrome is available in the United States only," and "only available in English."
• Your phone has to clear a hardware bar. At least 4GB of RAM, running Android 12 or higher. An older budget phone is excluded even in the US.
• You must be 18 or over on Android. On a computer the same feature starts at 13.
• Auto browse is the paid half. It needs Google AI Pro ($19.99/month) or Google AI Ultra (from $99.99/month) on a personal Google Account.
• Paid tasks are capped per day. Google publishes up to 20 requests a day on AI Pro and up to 200 a day on AI Ultra.
• You can walk away mid-task. Google says that after you start a task you do not need to stay in the Chrome app, and you get a notification when it finishes.
• iPhone gets the assistant but not auto browse. Google states the agent feature "is not available on iPhone or iPad devices."

Do You Get It? The Short Version

If you are an adult in the United States with an English-language Android phone from roughly the last four years, yes, and you pay nothing. Google's own requirements list is short and every line is a hard gate. Here is the whole thing.

The RAM and OS lines are the ones worth checking before you go looking for the button. Google names no phone models, only that floor, so the honest answer for a cheap or ageing handset is that it may simply never appear. Google also warns that the rollout is gradual: "Gemini in Chrome is not available to all users at this time."

What Does It Actually Do on a Phone?

It reads the page you are on and answers questions about it, without you leaving the tab. Google's announcement lists summarising long articles, answering questions about web pages, and connecting to Google apps "including Calendar and Keep" so you do not have to switch between tabs. It also generates and customises images with Nano Banana on the go.

Google's help page adds the everyday examples: summarise the key takeaways from an article, explain a complex topic another way, test your knowledge of something you are learning, modify a recipe to meet dietary needs, and compare information based on your preferences.

One desktop feature does not come with it. On a computer you can share up to 10 open tabs with Gemini at once. Google describes tab-sharing as a computer capability; on Android the assistant works from your current page.

What Is Auto Browse, and Do You Have to Pay?

Auto browse is the part that does things instead of telling you things, and yes, it is paid. Google's announcement is specific about who: "AI Pro and AI Ultra subscribers on Android in the U.S." Its examples are booking parking for an event, updating a recurring online order, and organising travel for a holiday.

The help page's own list is broader: comparing products and adding items to your cart, finding and booking travel accommodation, making restaurant reservations and scheduling appointments, drafting communications and organising calendar schedules, researching job markets, and "online finance chores."

Those daily caps are Google's own numbers and they matter more than they look. Twenty multi-step tasks a day on the $19.99 plan is plenty for errands and nowhere near enough to run anything continuously. Google says Gemini in Chrome tells you when you reach the limit, and that there is a separate cap on how many tasks can run at the same time.

Auto browse has its own requirements on top of the ones above: you must grant Chrome permission to auto browse on your behalf, hold the plan on a personal Google Account, and have Chrome's Safe Browsing set to Enhanced or Standard protection. It does not work in Incognito.

Android, Desktop and iPhone Are Not the Same Product

Three platforms, three sets of rules, and the differences are large enough to change the answer for you. This is the table to read if you use Chrome on more than one device.

Two consequences are worth stating plainly. A British Android user gets nothing today: the UK is on the desktop list but not the Android one. And a British iPhone user gets nothing either, because the UK is absent from the iPhone list as well. If you are in the UK, Gemini in Chrome on the desktop remains the only version available to you.

How to Turn It On, and How to Turn It Off

Open Chrome and tap Ask Gemini, or reach it from the three-dot menu at the top of the screen. Google's announcement names both routes: "tapping the Gemini icon or the three-dot menu on the top bar of your Android device." You opt in the first time you use it by following the on-screen instructions. Your current page is shared with Gemini by default, and you can tap Stop sharing in the text box at any time.

For auto browse, you describe the task, tap Submit, and Google shows you a plan before anything happens. Check that it understood your prompt and your personal details, then tap Start Task. While it runs there is an auto browse icon on the tab it is using, and you can tap Switch tabs to watch, Take over task to finish a step yourself, or Stop.

To switch the agent off entirely: Chrome, More, Settings, Gemini in Chrome, then turn off "Let Gemini browse for you." That leaves the ordinary assistant intact.

The Risks Google Publishes Itself

Google calls auto browse experimental and puts the responsibility on you, in writing. Its help page says: "You're responsible for Gemini's actions during a task, including mistakes and unexpected results like purchases." That sentence should decide how you use it.

The specific hazard it names is prompt injection -- hidden instructions on a page, in an email or in a document that the agent reads and obeys. Google's own examples of what a malicious instruction could do are taking private information from your emails or documents and posting it publicly, sending your Gmail to an outside service without you knowing, and exposing insights about you from your connected apps. The announcement says the models "are trained to detect known threats, such as prompt injection," and that auto browse "is designed to ask for confirmation before completing some sensitive tasks."

Google is also explicit that the agent may share your information with the sites it visits, and that it can use a broad range of websites "even if you haven't connected the related apps to Gemini Apps."

Should You Use It?

Turn on the free assistant. Summarising a long article, having a dense page explained back to you, or checking a recipe against a dietary restriction are low-risk, and you are not handing anything over to do it. If your phone qualifies, there is no argument against it.

Pay for auto browse only if you already pay for Google AI Pro. Do not buy a $19.99 subscription for this feature alone on a phone. Twenty tasks a day is a real cap, the feature is labelled experimental by its own maker, and everything it does you can still do yourself.

Use it for errands, not for money. Comparing products, checking availability, drafting a booking up to the point of payment: good. Anything that ends in a completed purchase, a financial account, or a message sent in your name deserves the confirmation screen and your full attention, because Google has said in advance that mistakes are yours.

If you are outside the United States, there is nothing to do. Not "not yet rolled out" -- Google's Android availability page names one country. Watch for it to change rather than hunting for a setting that is not there.

For the wider picture of what these agents are and how to think about letting one act for you, see our plain-English guide to what an AI agent actually is, and our explainer on Chrome auto browse on the desktop, which covers the version that uses your saved passwords.

The bottom line: the free half of this is a genuine upgrade to browsing on an Android phone and costs you nothing. The paid half is the most capable consumer AI agent you can put on a phone today, and Google has been unusually candid that it is experimental and that its mistakes are your problem. Check the 4GB and Android 12 lines first, because for a large number of phones this release changes nothing at all.

Sources: Google's announcement Tap into the power of Gemini in Chrome on Android, article:published_time 2026-08-18, for the US-wide Android availability, the Calendar and Keep connections, Nano Banana image generation, the AI Pro and AI Ultra gate on auto browse, the prompt-injection training claim, the confirmation step on sensitive tasks, and the two ways to open it. Google's Gemini in Chrome availability help article, read in its Android, Computer and iPhone & iPad variants, for the age, country, language and device requirements on each platform and for the locale lists. Google's Use Gemini in Chrome for the 4GB RAM and Android 12 floor, the opt-in flow and the examples. Google's auto browse help article for the Android auto browse requirements, the daily task caps, the leave-the-app behaviour, the off switch and the published risks. Prices from gemini.google/subscriptions, re-read August 18, 2026. All sources read on August 18, 2026.