Google's Gemini Spark Gets Full Chrome Browsing, Can Tap Logged-In Accounts and Saved Passwords

Google has rolled out Chrome web-browsing capabilities to Gemini Spark, allowing the AI assistant to use users' logged-in accounts and saved passwords to handle "tedious web errands," Engadget reported on August 3. The update turns Gemini from a chat-based assistant into something closer to a true browser agent.

What's New

Gemini Spark can now navigate the web on a user's behalf, complete forms, place online orders, and interact with web apps that require authentication. Crucially, it does so without forcing users to hand over credentials in a separate flow - Gemini uses the user's existing Chrome profile, including saved passwords and active sessions. The feature is rolling out first to Gemini Advanced subscribers in the US.

"This is the closest thing to an AI browser agent shipping at consumer scale," an Engadget analysis noted.

Why It Matters

Browser agents have been a 2026 theme. Anthropic launched Claude in Chrome earlier this year, and OpenAI's Operator launched as a research preview in late 2025. The difference with Gemini Spark is that it ships inside the browser users already use, with no extension to install. That gives Google a structural advantage in adoption, but it also concentrates risk: any Gemini Spark failure now has the keys to the user's entire logged-in web.

How Safe Is It

Google is rolling out Spark with a "supervised" mode that asks for user confirmation before any irreversible action - purchases, sent emails, deletions. A second "autonomous" mode is available for low-risk tasks, with a strict allow-list of sites. The company is publishing an explicit trust model in the Gemini app, and the launch comes a week before the EU's AI Act labeling rules reach full force - Gemini Spark is now labeled as an AI agent in line with the new requirements.