# Gemini in Chrome is now on every US Android phone, and it can browse for you > Google opened Gemini in Chrome to all US Android users on 18 Aug 2026, plus Auto Browse for Pro/Ultra subscribers. What it does, and when to use it. _Markdown version of https://gemini-notebook-hub.online/guides/gemini-in-chrome-android — for AI and LLM crawlers. Content is identical to the web page._ --- _7 min read · updated 19 Aug 2026_ Google flipped Gemini in Chrome on for all US Android users on 18 August 2026, and gave AI Pro/Ultra subscribers Auto Browse, an agentic mode that can book parking or update an order while you do something else. It's a genuinely different tool from Gemini Notebook, and mixing them up costs you either time or your research trail. --- > **Quick answer** > As of 18 August 2026, every US Android user with Chrome updated to the latest version can tap the Gemini icon (or use the three-dot menu) to summarize the open page, ask questions about it, and pull in Calendar or Keep, without switching tabs. AI Pro and AI Ultra subscribers additionally get **Auto Browse** on Android for the first time: an agentic mode that navigates sites and completes multi-step tasks like booking parking or updating a recurring order, pausing to ask you to confirm before anything sensitive like a payment. > It's not a replacement for Gemini Notebook. Chrome's Gemini gives you a disposable answer about whatever's on screen right now, logged into a general, time-limited activity feed. Gemini Notebook builds a permanent, citation-backed research library from sources you deliberately choose to keep. The rollout closes a gap that's existed since Gemini in Chrome reached general availability on desktop back in September 2025 (after debuting for AI Pro/Ultra subscribers at Google I/O in May 2025): Android users could see the feature mentioned everywhere and still not have it. Per [Google's own announcement](https://blog.google/products-and-platforms/products/chrome/gemini-in-chrome-android-auto-browse/), the wider release also brings Auto Browse to phones for the first time, a capability that's been available to AI Pro subscribers on desktop Chrome since late January 2026. ![Google's official blog post announcing Gemini in Chrome's expansion to all US Android users, with Auto Browse for AI Pro and AI Ultra subscribers](/blog/gemini-in-chrome-android-blog.png) _Google's announcement post, blog.google, 18 August 2026._ ## How to turn it on - Update Chrome to the latest version. You need Android 12 or later with at least 4GB of RAM, your device language set to English (US), and a Google account signed in for a user 18 or older, outside Incognito mode. - Tap the Gemini icon to the right of the address bar. Chrome only lets you pin one custom toolbar button, so if yours is set to something else, use the three-dot overflow menu and tap **Ask Gemini** instead. - The first time you use it, you'll get a one-time opt-in prompt. The icon shows up by default for eligible accounts, but actually using the feature requires that explicit consent step first. - A chat sheet slides up from the bottom. By default it shares your current tab's content for that conversation; a **Stop sharing** control in the input box lets you withhold the page per-session. - Work or school Workspace accounts need admin approval before the icon appears at all. What it can summarize has real limits worth knowing before you rely on it. Paywalled or logged-in pages are reported as unreliable, since the model only sees what's rendered in the page, not what's behind an authentication wall it can't pass. Chrome's PDF viewer got its own **Summarize with Gemini** button, but scanned or image-only PDFs without selectable text produce weaker summaries, the same OCR problem NotebookLM users already know from image-only sources. Desktop Chrome can summarize across up to 10 open tabs at once; nothing in Google's Android-specific coverage confirms that multi-tab summarization has reached the phone version yet, so treat it as single-page-only on Android until Google says otherwise. ## What Auto Browse actually does Auto Browse is the agentic layer on top of the summarizer, and it's gated behind an AI Pro or AI Ultra subscription. Google's own examples: booking parking for an event, updating a recurring online order, or organizing travel for a trip. It navigates across sites, clicks links, and fills forms on your behalf, and when permitted it can pull saved logins from Google Password Manager to sign in automatically along the way, a permission worth noticing before you grant it. > **Quick answer** > **The guardrail that matters most:** Auto Browse is designed to stop and ask for your explicit confirmation before the last step of anything sensitive, like posting publicly or actually paying for a purchase. It will build a cart or fill in a form on its own, but per Google's own description of the feature, it hands control back to you before money moves or something goes public. Google's stated defenses against the obvious risk, a malicious page tricking the agent into doing something you didn't ask for, include prompt-injection content classifiers, markdown sanitization, and Safe Browsing checks on suspicious URLs before the agent follows a link. That's a real mitigation layer, not a guarantee: in March 2026, security researchers at [Unit 42 disclosed CVE-2026-0628](https://unit42.paloaltonetworks.com/gemini-live-in-chrome-hijacking/), a flaw that let a low-privilege Chrome extension inject code into the desktop Gemini side panel and inherit its camera, microphone, and file-access permissions. Google patched it in January 2026, before the write-up went public, but it's a concrete reminder that the agentic surface in Chrome has already drawn real security scrutiny once. ## Gemini in Chrome vs. Gemini Notebook: which one to actually use These solve different problems, and it's easy to reach for the wrong one. Chrome's Gemini is built for a quick, disposable answer about whatever's open right now. Nothing in Google's product page, support docs, or the Auto Browse announcement mentions citations, source linking, or a saved artifact from a summarization session; the only history it points to is the same general **Gemini Apps activity** dashboard used across all of Gemini, not a per-topic or per-page record. That activity log auto-deletes on a timer (18 months by default, adjustable to 3, 18, or 36 months) unless you turn it off entirely, in which case nothing past about 72 hours is kept at all. Gemini Notebook is built for the opposite job. You deliberately add sources, they stay in the notebook indefinitely as a defined corpus, and every answer the chat gives you is grounded with an inline citation pointing back to the exact passage it came from. If you close the tab and come back in three months, the notebook and its sources are exactly where you left them. ``` Gemini in Chrome Gemini Notebook ------------------------------------------------------------------ Scope one page at a time a corpus you choose Persistence none (disposable chat) permanent Citations no yes, per-answer History generic activity log the notebook itself Multi-source synth up to 10 tabs (desktop) up to 300-600 sources Best for a quick answer now research you'll revisit ``` The practical rule: reach for Chrome's Gemini when you want a throwaway summary of the page in front of you and won't need to prove where a claim came from later. Save the page as a source in a Gemini Notebook the moment you know you'll want to combine it with other material, cite it, or come back to it. A reasonable workflow is to use Chrome's Gemini to triage a pile of tabs quickly, then commit only the ones that matter into a notebook for the real work. ## People also ask ### Is Gemini in Chrome free on Android? The core summarize-and-ask feature is free with any Google account. Auto Browse, the agentic task-completion mode, requires an AI Pro or AI Ultra subscription and is currently US-only. ### Does Gemini in Chrome save my page summaries anywhere I can find them later? Not as a distinct, organized artifact. It logs into the general Gemini Apps activity feed, which is account-wide and time-limited (18 months by default), not a per-page or per-topic record you can browse and cite later the way a Gemini Notebook source is. ### Can Auto Browse make purchases without asking me? No. Google's own description says it's designed to stop and ask for explicit confirmation before sensitive final steps like completing a purchase or posting publicly. It can build a cart or fill in a form on its own, but it hands control back to you before anything is finalized. ### Is Gemini in Chrome available on iPhone or desktop too? Yes. Desktop Chrome had it first, reaching all US Mac and Windows users in September 2025 after debuting for AI Pro/Ultra subscribers in May 2025. It's also in Chrome for iOS via the omnibox, though Google notes some agentic 'Skills' capabilities aren't yet planned for iOS the way they are on desktop and Android. ### Should I use Gemini in Chrome instead of Gemini Notebook for research? Use Chrome's Gemini for a quick answer about one page you're on right now. Use Gemini Notebook when you're building anything you'll revisit, combine with other sources, or need to cite, since it's the one built for permanent, sourced work rather than a disposable in-browser chat.