Google has officially rolled out Chrome’s new “auto browse” feature, powered by Gemini 3, promising to handle multi-step online chores on your behalf.
Available now for AI Pro and AI Ultra subscribers in the US, the upgrade transforms Chrome into a semi-autonomous assistant capable of scrolling, clicking, and filling forms while you continue browsing elsewhere.
Unveiled as a “powerful agentic experience” designed to reduce the friction of repetitive online tasks, auto browse is triggered by a prompt to Gemini in Chrome and opens a glowing tab marked with a cursor and sparkle icon, while a side panel shows step-by-step progress.
Actions are executed locally on your device, with cloud models leveraged for intelligence, ensuring sensitive steps such as auto-filling credentials via Google Password Manager remain under user control.
However, Gemini still requires manual confirmation for critical actions like pressing the buy or post button, maintaining a balance between automation and accountability.
The upgrade is not just about convenience; it’s about scale. Google demonstrated scenarios ranging from shopping with budget constraints to scheduling appointments, filing expense reports, and even registering sports teams.
For example, a user could instruct Chrome: “Go to Etsy and find supplies to recreate the Photo Booth and add them to my cart. Don’t spend more than $75 total.” Auto browse would then identify items, apply discount codes, and keep spending within the limit.
Similarly, prompts can handle travel planning by comparing hotel and flight costs across multiple weekends, offering the cheapest option overall.
The rollout comes amid intensifying competition in AI-driven browsing, with rivals like Microsoft Edge and Perplexity experimenting with agentic assistants. Chrome’s advantage lies in its deep integration with Google services, from Gmail and Calendar to Maps and Flights, making auto browse a natural extension of the ecosystem.
UK users will need to wait, as the feature is currently limited to US subscribers, though Google has a track record of expanding availability once initial testing stabilises.
Sadly, there are caveats. Auto browse has a daily limit on agentic actions, and while it can streamline chores, it is not fully autonomous. Users must remain engaged, authorising sensitive steps and retaining the ability to “Take over task” at any point.
Privacy concerns also linger, as semi-autonomous browsing inevitably raises questions about data handling, even with Google’s assurances that actions occur locally.
