What can Gemini in Chrome now do for Android users in the United States?

Google has rolled out Gemini in Chrome to all Android users in the United States. The feature turns the browser into a browsing assistant that summarizes articles, answers questions about the page, and links to Calendar and Keep without leaving the tab.

The rollout and how to reach it

Google announced the rollout on August 18, 2026. The feature arrives inside the Chrome app on Android phones across the United States. You open it by tapping the Gemini icon or the three-dot menu on the top bar. No separate download required.

Summaries, questions, and app links without leaving the page

Once opened the assistant reads the page you are viewing. It can condense a long article into a short summary by extracting key points from the visible text. You can also ask it questions about the content without copying text elsewhere because the context stays attached to the tab. Integration with Google Calendar and Google Keep works the same way, letting you create events or notes while the page remains visible so switching between apps becomes unnecessary.

Image generation

A model called Nano Banana powers picture creation inside the browser. You customize or generate images directly on the page. The announcement does not explain which model version powers Nano Banana or whether it runs locally or in the cloud.

Auto browse for subscribers

Subscribers to AI Pro and AI Ultra plans unlock a second layer called auto browse. This agentic mode performs multi-step tasks on your behalf. It books parking for an event, updates a recurring online order, or organizes travel for a vacation. The system asks for confirmation before it finishes sensitive actions. Models are trained to detect known threats such as prompt injection, and security is described as built in by design.

What the announcement leaves out

The announcement covers only Android users in the United States with no date given for other countries, iOS, or desktop Chrome. Pricing details for the subscription tiers are not included in the post. What happens when the assistant misunderstands a complex request is not described. The blog post does not share latency numbers or success rates for auto browse tasks. Users will discover the practical limits through daily use.