Google Introduces Auto Browse AI Agent for Chrome Users

Google has released Auto Browse, an AI-driven tool for Chrome that can handle complex web activities such as shopping, booking travel, and managing online forms. The rollout is part of a broader effort to integrate artificial intelligence into everyday browser functions for select users in the United States.

Auto Browse operates through the Gemini 3 model and is accessible to subscribers of Google AI Pro and AI Ultra plans. The tool is designed to automate multi-step online processes, including comparing products, applying discounts, and adding items to shopping carts. It can also log into websites using Google Password Manager, provided the user grants permission.

The feature is accessed via a fixed Gemini sidebar within Chrome, which replaces the earlier floating window interface. Users can initiate and monitor several Auto Browse tasks at once, and the browser marks tabs controlled by the AI with a distinct visual indicator. This allows users to continue their own browsing activities while the AI completes assigned tasks.

Auto Browse processes all web interactions in the cloud, streaming page content to the Gemini model for analysis and action. This approach means that browsing data is not handled locally on the user's device but is instead managed through Google’s cloud infrastructure.

What the numbers show

• Auto Browse launched in the U.S. on January 28, 2026
• Access is limited to Google AI Pro and AI Ultra subscribers
• Gemini 3 is the underlying model powering the feature

Before conducting sensitive actions such as making purchases or posting on social media, Auto Browse pauses and requests explicit user confirmation. This step is intended to ensure that users remain in control of important transactions and account activities initiated by the AI agent.

Security considerations have been identified, including potential vulnerabilities such as prompt injection attacks. There are also open questions about how Google manages and stores the web page data processed by Auto Browse, given that all activity is routed through cloud servers.

Auto Browse supports concurrent execution of multiple tasks, enabling users to delegate various web activities to the AI while continuing with other browsing. The persistent sidebar and visual indicators are intended to help users track and manage ongoing automated processes within Chrome.

* This article is based on publicly available information at the time of writing.