Chrome competes with AI browsers with tighter Gemini integration and proxy features for standalone tasks

2025 was the year in which a host of AI browsers were launched from companies like OpenAI, Perplexity, Opera, and The Browser Company with the aim of replacing Chrome with features like sidebar assistants and automated tasks. Now, Google is showcasing its capabilities in the field of artificial intelligence by adding similar features to Chrome, which is the largest browser in the world by market share.

While Google brought Gemini to Chrome last September, the assistant lived in a floating window. With this update, the company will put its AI assistant in a persistent sidebar, so you can ask questions about your current website or other open tabs.

One interesting feature that Google showed off to press ahead of today’s launch includes multiple tabs. When you open different tabs from a single web page, Gemini’s sidebar understands them as a contextual group. This is useful when comparing prices or different products you are considering purchasing.

Before today, Gemini in Chrome was only available to Windows and MacOS users. With this launch, the sidebar will be available to Chromebook Plus users as well.

Google is also taking advantage of its newly launched Personal Intelligence feature, which connects to your Gmail, Search, YouTube and Google Photos accounts, allowing you to ask questions based on your own data. This feature will be rolling out to Chrome over the coming months, meaning you can ask Gemini in the sidebar about things like your family schedule, or ask him to draft and send an email without switching to Gmail.

There’s a new Nano Banana integration coming to Chrome as well, which allows you to edit an existing photo with another photo or product you find while browsing.

However, the most ambitious feature is called Auto-Browsing, which aims to handle tasks for you using your personal information and browse websites for you. For example, you can ask the agent to go to a specific website, purchase an item for you, and find a discount coupon. The agent will request your intervention when performing data-sensitive tasks, such as logging into a website or making a final purchase.

Last year, the company explained that these features would use Chrome’s password manager or saved card details, but said its AI models would not be exposed to any of these details.

This feature is initially rolling out to AI Pro and Ultra subscribers in the US

Browser-based agents are difficult and often fail to complete tasks. Google’s demo, just like many other AI demos, involves shopping and travel planning. In real-world use cases, agents often don’t understand intent or get interrupted while traversing different locations, and this can pose a challenge for wider adoption.

In its early testing, users used the feature for tasks like scheduling appointments, filling out tedious online forms, collecting their tax documents, getting quotes for plumbers and electricians, and submitting expense reports, the company said.

Gemini sidebar support and Nano Banana integration will arrive starting today, while the Personal Intelligence feature will be available in “the coming months,” the company said.