🎧 Prefer to listen?

Your browser just got an upgrade that most people haven’t noticed yet. Google quietly turned Chrome into an AI assistant that can read your tabs, answer questions about what you’re looking at, and — if you’re paying for the premium tier — actually browse the web on your behalf. Shopping carts get filled. Reservations get booked. Research gets summarized while you do something else.

I tested what’s available right now. Here’s what’s free, what costs money, and whether you should actually trust your browser to do your browsing.

What Gemini in Chrome actually does (for free)

If you’re using Chrome on desktop or mobile, there’s a Gemini button you might not have noticed. It lets you ask questions about whatever page you’re currently on. Highlight a paragraph and ask it to explain. Compare prices across open tabs. Summarize a long article you don’t have time to read.

This is free and available now. No subscription needed.

On Android, it works with anything on your screen — not just Chrome. On iOS, Gemini is built right into the Chrome app. Either way, you can ask contextual questions about what you’re looking at without switching apps.

The AI Mode in Search is the bigger deal. You can now attach Chrome tabs, images, and files directly to your search query. Instead of typing “compare these three laptops,” you can literally show Search the three tabs you have open and ask it to compare them. This works right now, for free, wherever AI Mode is available.

I covered the broader Google I/O announcements in every free AI tool they just launched — Chrome is just one piece of a much bigger picture. If you’re comparing AI tools across platforms, the AI tools with the highest satisfaction rates is a good reference.

Auto Browse — the feature that shops for you

Here’s where it gets interesting. Google’s new Auto Browse feature puts Gemini in the driver’s seat. You tell it what you want — “find me a dinner reservation for two on Saturday in Brooklyn” or “add these items to my Target cart” — and it navigates the web, clicks buttons, and fills out forms on your behalf.

What’s available:

  • Available now for Google AI Ultra and Pro subscribers in the U.S. only.
  • Works for shopping, booking reservations, and similar multi-step web tasks.
  • Runs inside Chrome — no new app to install.

What it can’t do (yet):

  • Limited to U.S. sites and services.
  • Can’t handle sites with complex authentication or two-factor steps.
  • It makes mistakes — Google says to check responses for accuracy.

This is the first mainstream browser that can actually do things on the web for you, not just show you things. That’s a real shift from “search and click” to “tell and done.”

Gemini Spark — your 24/7 browsing agent

The more ambitious feature is Gemini Spark, which runs in the cloud 24/7. Unlike Auto Browse, which works while you’re watching, Spark keeps working after you close your laptop.

What’s available:

  • Beta launching next week for Google AI Ultra subscribers.
  • Runs on Gemini 3.5 Flash — Google’s fastest model, clocking 1,500 tokens per second.
  • Works in the cloud, not on your device.

How it’s different from Auto Browse:

  • Auto Browse = “do this while I watch”
  • Spark = “do this overnight and tell me when it’s done”

Think of it as the difference between cooking alongside someone and having a private chef who prepares everything while you sleep. Spark can research, compare, organize, and present findings without you being present.

For anyone managing multiple projects — which, if you’re reading this, probably includes you — this could replace the “I’ll research that later” pile that never actually gets done. I wrote about building your first automation in 15 minutes — Spark is basically that idea scaled to your entire browser.

SynthID — knowing what’s real

Chrome is also adding SynthID verification “over the coming weeks.” This is Google DeepMind’s AI watermarking technology that detects AI-generated images, videos, and audio.

For free, right in your browser, you’ll be able to check if content you’re looking at was generated by AI. OpenAI, ElevenLabs, and others are implementing SynthID into their outputs too.

This won’t solve deepfakes, but it’s the first time a major browser will have built-in AI detection at the consumer level. You don’t need to be technical. You just need Chrome.

One-click password changes

Small but useful: Chrome can now detect compromised passwords and offer to automatically change them for you. One click, AI handles the password change flow on the site, and Chrome saves the new password to Google Password Manager.

Available on supported sites, for free. If you’ve been putting off updating breached passwords, this removes the friction completely.

Should you actually use these features?

Here’s my honest take.

Free features — yes, start now. Attaching tabs to search, asking Gemini questions about pages, SynthID verification, and one-click password changes are all net-positive. You’re not giving up anything you weren’t already sharing with Google, and the utility is immediate.

Auto Browse — worth trying if you’re already paying for AI Pro/Ultra. The shopping and booking automation is genuinely useful for repetitive tasks. It’s not perfect, but it’s better than doing it manually for straightforward requests.

Gemini Spark — wait and see. The cloud-based agent concept is powerful, but “beta next week” means it’s going to be rough. Try it if you’re an Ultra subscriber, but don’t rely on it for anything important yet.

If you’re not using Chrome’s AI features at all, I’d start with the tab-attached search. It’s free, it’s immediate, and it’s the feature most people don’t know exists yet.

I compared browser-based AI tools across platforms in the privacy problem nobody talks about — worth a read if you’re concerned about what Chrome is learning while it helps you.

The bottom line

Chrome just went from being a window onto the web to being a participant on the web. The free features are genuinely useful today. The paid features are a preview of where browsing is headed — from “you do the work” to “you describe the work and watch it happen.”

Whether that’s exciting or terrifying probably depends on how much you trust Google. Either way, it’s happening. If you want to understand what’s happening under the hood, what is an LLM — explained like you’re 5 breaks down the engine powering all of this.

Want help picking which AI tools to actually use? Check out the AI Tool Advisor or head to Start Here for a guided walkthrough.