Particle.news
Download on the App Store

Google Builds Agentic Browsing Into Chrome With Gemini 3’s Auto Browse

A desktop rollout introduces a side panel plus on‑page image edits, with Auto Browse limited to U.S. AI Pro and Ultra subscribers.

Overview

  • Gemini now lives in a persistent Chrome side panel on Windows, macOS and Chromebook Plus, enabling multitasking like summarizing pages and comparing items across tabs.
  • Auto Browse can click, type, scroll and complete multi‑step tasks across sites, with a visible Pause or Take over control and prompts before high‑stakes actions such as purchases.
  • Connected Apps let Gemini use context and take actions in Gmail, Calendar, Maps, YouTube, Google Flights and Shopping, configurable via settings.
  • Commerce tasks are backed by Google’s Universal Commerce Protocol co‑developed with Etsy, Shopify, Target and Wayfair to streamline agent checkouts.
  • Google says Personal Intelligence will come to Chrome in the coming months as an opt‑in feature, while reporting flags ongoing privacy, reliability and website pushback concerns.