How to Use Gemini Auto Browse in Chrome on Android

Google rolled out Gemini in Chrome to Android users in India on 7 October 2026. At the same time, Auto Browse became available on both desktop and Android for Google AI Pro and Ultra subscribers in the country. The feature lets Gemini handle multi-step web tasks for you, such as filling forms, comparing options, or ordering groceries, while you review its plan first.

Auto Browse runs in its own tab and pauses for confirmation on sensitive actions. This guide covers the requirements, the exact steps on Android, practical prompts, daily limits, and ways to stay in control.

Requirements for Auto Browse on Android

You need a Google AI Pro or Ultra subscription on a personal Google Account. The free Gemini plan does not include Auto Browse. You must be at least 18 years old. The device language must be set to English (US or UK) at launch in India, with more languages planned. Use the latest version of Chrome and enable Standard or Enhanced Safe Browsing. Auto Browse does not work in Incognito mode or on school accounts.

Gemini in Chrome itself (summaries, page questions, Nano Banana image tools, and connections to Keep or Calendar) is available more broadly to Android users in India, even without a paid plan. Auto Browse is the paid agentic layer.

If you already use Gemini models for other work, see the Gemini 4 Argon access and pricing guide for how Pro and Ultra plans fit together.

Start an Auto Browse task on Android

Open Chrome and sign in with the eligible account. Look for the Ask Gemini control, usually at the bottom or top of the browser. Tap it to open the panel.

Describe one clear multi-step task. State the outcome, constraints, and what you do not want done. Examples:

• “Compare three hotels in Mumbai for 12–15 November under ₹12,000 per night with free cancellation. Shortlist the two best and do not book.”
• “Find parking near the stadium for Saturday evening and fill the form with my saved details, then stop before payment.”
• “Order weekly groceries from my usual store. Add the items from this recipe and show me the cart before checkout.”

Submit the request. Gemini shows a plan that lists the sites it will visit and the information it may share. Read the plan. If it matches what you want, tap Start Task. Gemini opens a dedicated tab marked with an Auto Browse indicator and begins working. You can switch to that tab to watch, leave the app, and return later.

You can take over a step at any time, then resume or hand the task back. Gemini asks for confirmation before certain sensitive actions such as completing a payment, accepting terms, or creating an account. With permission, it can sign in via Google Password Manager without seeing the passwords.

Daily limits and what counts as a task

AI Pro subscribers receive up to 20 multi-step Auto Browse tasks per day. AI Ultra subscribers receive up to 200. Chrome also limits how many tasks can run at once. When you hit a limit, the interface tells you. A task that only summarizes a page or answers a question does not count against the Auto Browse quota; those features are available more widely.

The feature is experimental. Google trains the models to detect threats such as prompt injection. Still, review every plan and never let it submit payment details without your explicit check.

Useful prompts and tips for better results

Be specific about sites when you know them. “Go to the REI site and add camping gear for this photo” works better than a vague “find gear.” Mention constraints such as price caps, free cancellation, or “do not submit the form.”

Start with research or comparison tasks before you try anything that spends money. After Gemini builds the cart or shortlist, review it yourself and finish the sensitive step manually.

Connect Google apps if you want context from Calendar or Keep, but keep the first task simple so you can see how the plan looks. Update Chrome if the Ask Gemini control is missing; rollouts are staged.

Safety and control steps

Always read the plan before you tap Start Task. Note which sites will be visited and what personal details may be shared. Take over immediately if the agent heads to an unexpected site.

Do not paste passwords, card numbers, or health information into the chat. Use Password Manager for logins when offered. Stop the task from the Gemini panel if anything looks wrong. Because the agent works in a separate tab, you can close that tab to end the run.

Enhanced Safe Browsing helps, but it does not replace your review. For users in India, the English-language requirement at launch means non-English device settings may not show Auto Browse yet.

What Auto Browse does not do

It does not replace manual checkout for every purchase. Sensitive final steps still need your confirmation. It is not available in Incognito or on school accounts. Free-tier users get the rest of Gemini in Chrome (summaries, questions, Nano Banana) but not the agent that navigates and fills forms across sites.

Desktop and Android share the same plan requirements and daily caps in India. Availability outside the announced regions remains limited; check the Ask Gemini panel on your device.

Bottom line

Auto Browse in Gemini in Chrome on Android lets Pro and Ultra subscribers offload multi-step web chores after they review a plan. Open Ask Gemini, describe a concrete task with clear boundaries, check the proposed sites and data, then start the run. Stay within the daily limits (20 for Pro, 200 for Ultra), watch for confirmation prompts, and finish payments yourself. The broader Gemini in Chrome tools remain available to more Android users in India even without a paid plan.

Sources

• Bringing Gemini in Chrome with auto browse to India — Google, 7 October 2026
• Gemini in Chrome, your AI browsing assistant — Google Chrome on YouTube
• Chrome Help documentation and coverage confirming Pro/Ultra requirements, 20/200 daily task limits, English language, and opt-in plan review for India Android and desktop