Google did not build a new browser for its agent. It put Gemini inside Chrome, gave it its own tab, and kept the irreversible clicks with the person. For Chrome-standardized companies that is the shortest path to a governed pilot.
Key Takeaways
- An agent in the browser you already manage. — Auto browse, launched January 28, 2026, lets Gemini scroll, click and type in a separate, marked tab while you keep browsing.
- The final click stays with you. — Gemini prepares purchases and social posts but you press buy or post; password autofill needs your approval; a daily action limit applies.
- Narrow availability. — Auto browse needs Google AI Pro ($20/mo) or AI Ultra ($100/mo) and is US only. It reached Android in late June 2026.
What Gemini in Chrome does
Gemini sits in a Chrome side panel as an assistant. Auto browse, added on January 28, 2026, turns it into an agent: give it a task, and it opens a tab marked with a cursor icon, then scrolls, clicks and types until it is done. Chrome shows an indicator and a glow around the tab while the agent works, and you can take over at any point.
The design is conservative on purpose. Gemini prepares a purchase or a social post, but you press the button. Password Manager autofill needs your approval, and a daily cap limits agentic actions. Auto browse reached Android in late June 2026 (nohacks.co).
Scorecard: 15/20
| Criterion | Score | Why |
|---|---|---|
| Agent | 4/5 | Multi-step tasks in a background tab; capped by daily limits. |
| Control | 4/5 | Managed through existing Chrome Enterprise tooling. |
| Safety | 4/5 | Human checkpoints on purchases, posts and autofill; visible agent tab. |
| Reach | 3/5 | US only, paid Google AI plan required. |
Pricing and availability
Auto browse is for US subscribers to Google AI Pro, about $20 a month, or AI Ultra, which Google cut from $249.99 to $100 a month at I/O in May 2026 (Framia). There is no separate browser to deploy: it ships in Chrome, so existing Chrome Enterprise management applies. Check Google's current enterprise policy list for the Gemini controls you need before a rollout.
Who it fits, and who should look elsewhere
- Fits: US companies standardized on Chrome and Google Workspace that want an agent pilot with human checkpoints on anything irreversible.
- Look elsewhere: outside the US, if you need domain-level limits on agent actions documented today (Comet Enterprise, Claude Team and Enterprise), or if daily action caps would block the workflow.
Gemini in Chrome alternatives
- Perplexity Comet: a standalone agent browser on every platform.
- Claude in Chrome: an extension with a safety classifier and admin site lists.
- Microsoft Edge, Opera Neon and Dia: see the full ranking.
Related questions
- Which AI browsers are worth shortlisting in 2026?
- Comet, Gemini in Chrome or Claude in Chrome: which fits?
- Should a company block, sandbox or pilot agentic browsers? (partner site CTAIO)
- How do you show up in Google AI Overviews?
Sources
- 9to5Google, Chrome Gemini auto browse (Jan 28, 2026).
- Framia, Google cuts AI Ultra to $100 a month (May 2026).
- nohacks.co, The agentic browser landscape in 2026 (updated Sep 27, 2026).
What is auto browse in Chrome?
Auto browse is Gemini's agent mode in Chrome. You give Gemini a task, and it opens a tab marked with a cursor icon, then scrolls, clicks and types to complete it while showing its steps in the side panel. You can take over at any time.
How much does Gemini in Chrome cost?
The Gemini assistant in Chrome is widely available, but auto browse needs a Google AI Pro subscription at about $20 a month or AI Ultra, which Google cut to $100 a month in May 2026. It is limited to the US.
Can Gemini in Chrome buy things for me?
It can do the work up to the purchase, but Google requires you to press the buy button yourself, and the same applies to posting on social media. Password autofill also needs your authorization.
Ready to Find the Right AI Tools?
Browse our data-driven rankings to find the best AI tools for your team.