Gemini in Chrome Guide (2026): Auto Browse, Side Panel and Skills

Chrome is no longer just a browser with a search box. Gemini now sits at the top of the window, opens in a side panel next to any page, can read across your open tabs, and with auto browse it can click through websites to finish tasks for you. Google rolled out the side panel and auto browse in January 2026, added Skills and expanded Gemini in Chrome to Android after Google I/O in May. This guide shows what each feature does, who can use it, how to start a task safely, and how to turn it all off if you prefer a plain browser.

TL;DR

Gemini in Chrome (side panel chat that understands your tabs) is available on desktop Chrome and, since late June 2026, on Android devices with 4 GB+ RAM set to English (US).
Auto browse runs multi-step tasks on websites. It needs a Google AI Pro (20 tasks/day) or AI Ultra (200 tasks/day) plan, a personal account, age 18+, the US, and English.
Start a task: click Ask Gemini at the top of Chrome, describe the task, review the plan, click Start Task.
Gemini asks before sending messages, submitting forms or buying; payments and accepting terms are always left to you.
Skills save multi-tab prompt workflows you can rerun with one click.

What Gemini in Chrome Can Do

Ask about the page: summarise an article, explain a technical doc, pull key dates from a policy page.
Work across tabs: compare three laptops open in different tabs and build a table of specs and prices.
Connect Google apps: add an event to Calendar, save a note to Keep or find something in Gmail without leaving the page.
Auto browse: let Gemini navigate sites, fill in forms and complete multi-step tasks like finding in-stock items or booking an appointment.
Skills: save reusable prompt workflows.
Pointer prompting: select part of a page or an image with your cursor and ask about just that bit.
Nano Banana on Android: create or edit images while browsing, such as turning a page into an infographic.

Requirements and Availability

Feature	Who gets it	Platforms
Gemini side panel	Signed-in users in supported countries and languages (US English first)	Windows, macOS, ChromeOS; Android since late June 2026
Auto browse	Google AI Pro or Ultra, personal account, 18+, US, English	Desktop; Android announced at I/O 2026
Skills	Gemini in Chrome users	Desktop
Nano Banana in Chrome	Gemini in Chrome users	Android

Auto browse does not work in Incognito, with school accounts, or on iPhone and iPad. Work accounts need an administrator to allow it. Chrome’s Safe Browsing must be set to Standard or Enhanced protection.

How to Use the Gemini Side Panel

Update Chrome: three-dot menu > Help > About Google Chrome.
Sign in to Chrome with your personal Google account.
Click Ask Gemini (the Gemini button at the top of the browser). The panel opens on the right, beside the page.
Type or speak your question. Gemini uses the current page as context by default; you can add other open tabs to the conversation.
Use the suggested chips (Summarise, Explain, Key takeaways) for one-click actions.

The panel stays open as you move between tabs, so you can keep researching while Gemini keeps context. That is its biggest advantage over copy-pasting into a chatbot.

An AI browser agent completing a booking task

How to Use Auto Browse (Step by Step)

Click Ask Gemini at the top of Chrome.
Describe the task clearly, for example: Find a hotel family room in San Diego for December 20 to 22 under $200 a night with free cancellation, and open the three best options.
Click Submit. Gemini shows a plan of the steps it will take.
Review the plan and click Start Task. Gemini works in a task tab you can watch.
When it needs you (a confirmation, a login or payment), it pauses and asks.
To stop, click Stop in the side panel, click Take over task on the task tab, or close the tab.

Limits: up to 20 multi-step tasks per day on AI Pro and 200 on AI Ultra, with a cap on how many run at once. Google says desktop auto browse will integrate with Gemini Spark in the coming months.

Safety: What Gemini Asks Before Doing

Google built several guard rails into auto browse:

Confirmation requests before starting a task, sending communications, modifying data, submitting web forms, scheduling events, or visiting sites with highly sensitive financial or health data.
Required takeover for finalising payments, accepting terms of service and creating accounts. Gemini will not do these for you.
Password permission: Gemini must ask before using Google Password Manager to sign in somewhere.

Browser agents also face prompt injection: a malicious page can hide instructions meant to trick the AI. My advice is the same as for any agent: avoid letting it loose on banking and government portals, watch the task tab on unfamiliar sites, and read every confirmation before you approve it.

Skills in Chrome

Announced at Google I/O 2026, Skills let you save a multi-tab prompt workflow and rerun it with one click. Examples:

“Compare the products in my open tabs and make a table of price, rating and warranty.”
“Summarise today’s open news tabs into five bullets each with the source.”
“Extract all job requirements from these listings and match them to my CV in Drive.”

Create one from a prompt you have just run (look for the save option in the side panel), give it a name, and it appears as a shortcut in Gemini in Chrome.

Gemini in Chrome on Android

Google launched Gemini in Chrome for Android in late June 2026 on devices with at least 4 GB of RAM and English (US) as the language. It summarises articles, answers questions about the page and connects to Calendar, Keep and Gmail. Personal Intelligence, which personalises answers with your Google data, is optional. Auto browse and Nano Banana image creation were announced for Android at the same time. On Android phones, Gemini has also replaced Google Assistant system-wide; see my Gemini replaces Google Assistant guide.

15 Prompts to Try

“Summarise this article in five bullet points and give me the main counter-argument.”
“Explain this page like I’m new to the topic.”
“Compare the phones in my open tabs: price, battery, camera, update policy.”
“Turn this recipe into a shopping list and add it to Keep.”
“Add this event to my calendar with a reminder the day before.”
“Find the cancellation policy on this site.”
“What do the reviews on this page complain about most?”
“Translate this page’s key points into Hindi.”
“Pull every date and deadline from this PDF.”
“Draft a reply to this forum post politely disagreeing.”
Auto browse: “Find a dentist near me with Saturday appointments and open the booking page.”
Auto browse: “Check which of these three stores has the item in stock in size M.”
Auto browse: “Fill this contact form with my details and stop before submitting.”
“Make an infographic of this page.” (Nano Banana on Android)
“Save this workflow as a Skill called Weekly price check.”

Gemini in Chrome vs ChatGPT and Claude Extensions

	Gemini in Chrome	ChatGPT for Chrome	Claude in Chrome
Built in	Yes	Extension	Extension
Agent tasks	Auto browse (AI Pro/Ultra)	Work mode	Yes
Google app integration	Calendar, Keep, Gmail	Via connectors	Via connectors
Mobile	Android	No	No

I covered OpenAI’s version in depth in my ChatGPT Chrome extension guide. If you live in Google’s apps, Gemini is the most convenient; if you already pay for ChatGPT or Claude, their extensions are capable alternatives.

Privacy Settings and How to Turn It Off

Gemini in Chrome sends page content you ask about to Google to generate answers, and your chats are saved in Gemini Apps Activity if that is on. Manage or delete them at myactivity.google.com/product/gemini.

To hide or disable Gemini in Chrome, open Settings in Chrome and look for the AI or Gemini in Chrome section, then turn off the toolbar button and features you do not want. Managed devices can be controlled by policy. If you are clearing AI from Windows too, see how to turn off Copilot in Windows 11, including Copilot in Edge.

What Gemini in Chrome Costs

Plan	US price	Side panel (ask about a page)	Auto browse tasks
Free Google account	$0	Yes, in supported countries	No
Google AI Pro	$19.99/month	Yes, higher limits	Up to 20 per day (US, 18+)
Google AI Ultra	$249.99/month	Yes, highest limits	Up to 200 per day (US, 18+)

For most people the free side panel covers the daily stuff: summarising articles, comparing tabs and explaining a page. Pay for AI Pro only if you will actually hand off errands such as filling forms, booking or shopping research several times a week. AI Pro also includes 2 TB of Google storage and Gemini in Gmail and Docs, which is what makes it good value for many people.

Worked Example: Planning a Weekend Trip

Open three tabs: a hotel booking site, a train or flight search, and a restaurant guide for the city.
Click Ask Gemini and type: Compare the hotels in my open tabs under $150 a night within 1 km of the old town, and list the three with the best reviews.
Gemini reads the open tabs and returns a short table. Ask follow-ups such as Which one has free cancellation?
With AI Pro, switch to auto browse: Check train times from my city on Friday after 5 pm and add the cheapest option to my Google Calendar. Gemini shows each step and pauses before anything that needs sign-in or payment.
Review the result, then book yourself. I recommend you always do the final payment step by hand.

Not Working? Fixes

No Ask Gemini button: update Chrome, sign in with a personal account, set Chrome’s language to English (US), and check you are in a supported country.
Auto browse missing: you need AI Pro or Ultra on a personal account in the US, aged 18+. It does not work in Incognito or with school accounts.
Task stops halfway: sites with CAPTCHAs or unusual layouts can block agents. Click Take over task and finish manually.
Hit the daily limit: 20 tasks on Pro, 200 on Ultra; limits reset daily.
Android panel missing: your phone needs 4 GB+ RAM and English (US).

Frequently Asked Questions

Is Gemini in Chrome free?

The side panel assistant is free for eligible users. Auto browse requires a Google AI Pro or Ultra subscription.

Is auto browse available outside the US?

Not yet. Google’s help page lists the US, English and age 18+ as requirements.

Can Gemini buy things for me?

It can find products and fill the cart, but you must complete the payment yourself.

Does Gemini in Chrome work on iPhone?

Auto browse is not available on iPhone or iPad. Use the Gemini app on iOS instead.

Which model powers it?

Google uses its current Gemini models; auto browse launched on Gemini 3-generation models in January 2026 and has been upgraded since.

Does Gemini in Chrome see my passwords?

Gemini reads the page content you ask about, and auto browse hands control back to you for sign-in and payment steps. Do not paste passwords into the chat, and avoid using it on banking or health pages.

Can I use Gemini in Chrome at work?

Only if your Google Workspace or Chrome Enterprise administrator allows it. Managed devices often have it turned off by policy.

Final Thoughts

The side panel is the feature most people will use daily, and it is free. Auto browse is genuinely impressive for tedious comparison and booking tasks, but treat it like a new assistant: give clear instructions, watch the first few runs and keep the final click for yourself. Skills are the hidden gem for anyone who repeats the same research every week.

For AI inside your documents, read my Copilot in Word, Excel and PowerPoint guide.

Gemini in Chrome Guide (2026): How to Use Auto Browse, the Side Panel and Skills