How to Use Gemini Auto Browse in Chrome on Android
Step-by-step guide to using Gemini Auto Browse in Chrome on Android for multi-step web tasks with Google AI Pro or Ultra.
Gemini Auto Browse in Chrome handles multi-step web tasks on your behalf. It opens sites, fills forms, compares options, and pauses for your confirmation before sensitive actions.
Google expanded the feature to Android users in India on 7 October 2026 for Google AI Pro and AI Ultra subscribers. Gemini in Chrome itself is available more broadly for summaries and questions about pages.
This guide covers requirements, the exact steps to start a task, and the safety checks Google built in. It draws from the official Gemini Apps Help page and Google’s India announcement.
What Auto Browse can do
Auto Browse completes chores that normally require switching between tabs. Official examples include booking parking for an event, filling a long form, ordering groceries, comparing products, and starting travel bookings.
It shows a plan first. You review which sites it will visit and what personal details it may share. Only after you tap Start Task does it begin. It runs in its own tab marked with an Auto Browse indicator.
You stay in control. You can take over at any point. It stops before final payments, accepting terms, or creating accounts. It can use Google Password Manager with your permission, but never sees your passwords.
Daily limits apply. Google AI Pro allows up to 20 multi-step requests per day. AI Ultra allows up to 200. It does not work in Incognito mode.
If you already use Skills for repeated instructions, Auto Browse pairs with them for web actions. See the guide to stacking Gemini Skills for related workflow tips.
Requirements before you start
Confirm these items so the feature appears.
- Android phone with the latest Google Chrome.
- Signed in to Chrome with a personal Google Account (not a school account).
- Google AI Pro or Google AI Ultra subscription active.
- Device language set to English (initially for Auto Browse in India; broader language support is planned).
- Age 18 or over.
- Safe Browsing set to Standard or Enhanced.
- Opted in to Gemini in Chrome.
Availability is rolling out. If the option is missing, update Chrome, restart the browser, and check again. Work or school accounts need administrator approval.
Gemini in Chrome for basic page summaries and questions is available to more Android users in India without a paid plan. Auto Browse specifically requires Pro or Ultra.
Start your first Auto Browse task
Follow these steps on Android Chrome.
- Open Google Chrome on your Android device.
- Tap the Ask Gemini icon (usually at the top or bottom of the browser).
- Describe a clear multi-step task. Example: “Find parking near the concert venue for Friday evening and fill the form with my contact details. Do not submit payment.”
- Review the plan Gemini shows. It lists intended sites and information it may share.
- If the plan looks correct, tap Start Task.
- Watch the dedicated Auto Browse tab. It shows progress. You can leave Chrome and return later.
- Approve any confirmation prompts for sensitive steps. Take over if needed for payments or final submissions.
On desktop the flow is similar: click Ask Gemini in the toolbar, describe the task, review the plan, then Start Task.
Be specific in prompts. Include constraints such as budget, date range, or “do not purchase.” Vague requests lead to broader plans that take longer to review.
Manage an active task
While Auto Browse runs, you have several controls.
- Switch to the task tab from the Gemini panel.
- Tap Take over task on the active tab to complete a step yourself.
- Resume or Give back task when you finish your part.
- Stop the task from the Gemini panel if you want to cancel.
You receive a notification when the task needs your attention. Sensitive actions always pause for confirmation. Google designed these stops to reduce the risk of unintended purchases or form submissions.
After the task, review the results carefully. Auto Browse can make mistakes, such as selecting the wrong option or missing a detail. Confirm anything that involves money, personal data, or legal agreements.
Safety notes from Google
Auto Browse is experimental. Google lists risks including prompt injection (malicious instructions hidden on a site), sharing personal information with websites, and possible errors in following instructions.
Safeguards include the plan review step, pauses before sensitive actions, and the ability to take over or stop at any time. You remain responsible for the actions taken.
Do not use it for high-stakes tasks without supervision. Avoid sharing sensitive financial or health data in prompts unless necessary, and review every plan.
You can manage permissions under Chrome Settings → AI innovations → Gemini in Chrome. There you can review sites it can sign into and turn the browse-for-you option on or off.
Tips for better results
- Start with low-risk research tasks before forms or bookings.
- Name specific websites when you know the best source.
- Include “stop and ask me before” language for any decision points.
- Keep Chrome updated and signed in with the same personal account.
- Check daily limits if the feature stops responding.
- Pair with Connected Apps (Calendar, Keep) for context when relevant.
Test the feature on a simple grocery list or product comparison first. Once the flow feels reliable, move to multi-site tasks.
Watch a demonstration of Gemini in Chrome features
This video walks through Gemini in Chrome updates, including Auto Browse examples of multi-step tasks.
Conclusion
Gemini Auto Browse turns Chrome into an agent that can complete everyday web chores after you approve a plan. On Android in India it is available now for Pro and Ultra subscribers, with Gemini in Chrome itself reaching more users for simpler help.
Review the plan, stay available for confirmations, and treat the results as a starting point rather than a final decision. Combined with Skills for repeated instructions, it reduces the time spent switching tabs.
Sources
Share this story
