The Auto Browse agent is available to AI Pro and AI Ultra subscribers, but there are some limits.
Google began stuffing Gemini into its dominant Chrome browser several months ago, and today the AI is expanding its capabilities considerably. Google says the chatbot will be easier to access and connect to more Google services, but the biggest change is the addition of Google’s autonomous browsing agent, which it has dubbed Auto Browse. Similar to tools like OpenAI Atlas, Auto Browse can handle tedious tasks in Chrome so you don’t have to.
The newly unveiled Gemini features in Chrome are accessible from the omnipresent AI button that has been lurking at the top of the window for the last few months. Initially, that button only opened Gemini in a pop-up window, but Google now says it will default to a split-screen or “Sidepanel” view. Google confirmed the update began rolling out over the past week, so you may already have it.
You can still pop Gemini out into a floating window, but the split-view gives Gemini more room to breathe while manipulating a page with AI. This is also helpful when calling other apps in the Chrome implementation of Gemini. The chatbot can now access Gmail, Calendar, YouTube, Maps, Google Shopping, and Google Flights right from the Chrome window. Google technically added this feature around the middle of January, but it’s only talking about it now.
Sidepanel with Gmail integration
Gemini in Chrome can now also access and edit images with Nano Banana, so you don’t have to download and re-upload them to Gemini in another location. Just open the image from the web and type in the Sidepanel with a description of the edits you want. Like in the Gemini app, you can choose between the slower but higher-quality Pro model and the faster standard one.
Can’t someone else do it?
Chrome’s new browsing agent (in preview) is the star of today’s show. The promise of agentic AI is one of the freedom to be lazy. Rather than filling out forms or copying information from emails, you simply task a robot with the job and relax while it goes to work. At least, that’s how it’s supposed to go. Most of the computer-use agents we’ve seen so far have been sluggish and unreliable enough to require human supervision. That kind of defeats the purpose. Now it’s Google’s turn to show us what its browsing agent can do.
Google says Auto Browse is based on its latest Gemini 3 models, with input from the company’s work on the experimental Project Mariner agent. If you’d do something with a keyboard and mouse inside your browser, Auto Browse is theoretically able to take over. It can also access the content and tabs in your browser, asking for permission when it needs sensitive data, such as your passwords.
Apartment hunting with Auto Browse
When you launch an Auto Browse task, Chrome will open new tabs as necessary and mark them with a sparkly AI icon so you know where the robot is active. You don’t have to keep that tab in the foreground, and you can even have multiple AI tasks going at once. The AI will ping you to check in when the task is done or when it needs your input. However, there are limits on Auto Browse similar to some of the company’s more computationally intensive chatbot functions. AI Pro subscribers get 20 AI browsing tasks per day, but those paying for AI Ultra get 200 per day.
Having the AI churning in the background may be a concern—after all, generative AI can make mistakes, and you are giving it complete control over the browsing experience. Google says it has implemented a range of security and safety rules to prevent misuse. For example, if you ask Auto Browse to research and buy something, it won’t actually buy it. Instead, it will find the item (hopefully) and progress to the purchase screen before letting you pull the trigger manually.
Filling a form with Auto Browse
Another thing to keep in mind is that Auto Browse doesn’t run locally. All content from your robotically operated tab is streamed to a cloud-based Gemini model. That means page content will be shared with Google, but the extent is not entirely clear. Google says Auto Browse is governed by the established Gemini in Chrome policy, which says Google stores information from websites in the Gemini Apps Activity (if Keep Activity is enabled). Page content is also “logged to your Google Account temporarily.” We asked if Google would use page contents processed by Auto Browse to further train AI models, but a spokesperson declined to provide specifics.
Chrome Auto Browse is rolling out today in preview, but it’s not exactly free. It won’t cost anything extra, but the feature is limited to AI Pro and AI Ultra subscribers. Google’s language suggests the feature may come to free users after the preview phase, but we’d expect strict usage limits.
Ryan Whitwam is a senior technology reporter at Ars Technica, covering the ways Google, AI, and mobile technology continue to change the world. Over his 20-year career, he’s written for Android Police, ExtremeTech, Wirecutter, NY Times, and more. He has reviewed more phones than most people will ever own. You can follow him on Bluesky, where you will see photos of his dozens of mechanical keyboards.
