Google Rolls Out Auto Browse Agent in Chrome Using Gemini 3
Google confirmed Auto Browse began rolling out in preview to AI Pro and AI Ultra subscribers in the U.S. and uses Gemini 3 to perform multistep web tasks.
Overview
Google confirmed it began rolling out Auto Browse, an autonomous browsing agent powered by Gemini 3, in preview to AI Pro and AI Ultra subscribers in the U.S.
The update moves Gemini into a persistent Chrome side panel and connects the assistant to Gmail, Calendar, YouTube, Maps, Google Flights and Shopping to perform multistep tasks across tabs, Google said.
Charmaine D'Silva, product lead for Chrome, demonstrated Auto Browse and said it will pause for user approval before completing sensitive actions such as purchases, according to a press demonstration.
Auto Browse runs in the cloud and is limited to 20 AI browsing tasks per day for AI Pro subscribers and 200 per day for AI Ultra subscribers, Google confirmed.
Google said Personal Intelligence will arrive in Chrome in the coming months, while a spokesperson declined to say whether page content processed by Auto Browse will be used to train models, leaving privacy disputes unresolved.
Analysis
Center-leaning sources frame Google’s Chrome AI update as a competitive, feature-driven advancement, using promotional language ("flexing its own AI muscle", "most ambitious") and prioritizing company demos and claims. Editorial choices spotlight convenience and rollout, while skepticism (agent fragility, privacy concerns) appears briefly and lacks independent expert voices.
Sources (5)
FAQ
Auto Browse is an autonomous AI agent powered by Gemini 3 that completes multistep web tasks like booking flights, shopping online, filing expenses, filling forms, researching travel options, managing subscriptions, and identifying items from images to add to carts within a budget.
It is available in preview to AI Pro and AI Ultra subscribers in the U.S., with limits of 20 tasks per day for AI Pro and 200 for AI Ultra.
Auto Browse pauses for user approval before completing sensitive actions such as purchases or sign-ins.
Personal Intelligence is an opt-in feature that allows Gemini to remember context from past interactions for more tailored responses, connecting to apps like Gmail and Calendar; it will arrive in Chrome in the coming months.
Google declined to say whether page content processed by Auto Browse will be used to train models, leaving privacy concerns unresolved.
History
This story does not have any previous versions.