Gemini In Google Chrome Now Live On Android In India, Auto Browse Rolls Out To AI Pro & Ultra Subscribers

Google has expanded Gemini in Chrome to Android users across India, months after bringing the AI browsing assistant to desktop in the country. Alongside this, the company has also switched on auto browse, Gemini's agentic feature capable of completing tasks on a user's behalf, for both desktop and Android, though access here is restricted to subscribers on Google's AI Pro and AI Ultra plans.

What does Gemini in Chrome actually do?

On Android, Gemini in Chrome functions as an in-browser assistant that can summarise long articles, answer questions about the page a user is currently viewing, and pull in other Google apps like Keep and Calendar without requiring users to switch between tabs. It also supports generating and customising images on the go using Google's Nano Banana image tool. For now, the Android version is limited to devices set to US-English or UK-English, with Google saying support for additional languages is planned.

What Auto Browse adds on top

Auto browse goes a step further than the assistant-style features above, by actually carrying out multi-step tasks inside the browser rather than just answering questions. According to Google, this includes things like booking parking for an event, filling out a lengthy form, or placing a grocery order for the week, tasks that would otherwise require manual back-and-forth across multiple web pages. Like the standard Android rollout, auto browse is currently limited to devices set to English, with more languages expected in future updates.

How to access it

Gemini in Chrome itself is rolling out to all Android users in India at no additional cost beyond having the Chrome app installed. Auto browse, however, is gated behind a subscription: it is available only to users on Google's AI Pro or AI Ultra plans in India, on both desktop and Android. Google has not specified exact in-app navigation steps in its announcement, so where the feature appears within Chrome's interface may vary slightly as the rollout reaches individual users.

Built-in safeguards for Agentic tasks

Given that auto browse can take real-world actions like placing orders or submitting forms, Google says the feature has been built to be secure by design, with its models trained to detect known threats such as prompt injection attacks, where malicious instructions hidden on a webpage attempt to hijack an AI agent's behaviour. As an added layer of control, Google says auto browse is designed to pause and explicitly ask users for confirmation before completing certain sensitive tasks, rather than acting entirely autonomously.