Google is continuing to expand its AI agent , Gemini Spark. Since July 30, 2026, Spark has been able to directly control the Chrome browser on desktop devices, using logged-in accounts and saved passwords to handle tedious online tasks on its own. At the same time, Google is making the agent available to Google AI Pro subscribers in over 160 additional countries. The catch for us: Germany and the entire European Economic Area are still not on the list.
Key Points at a Glance
- Starting July 30, 2026, Gemini Spark will be able to control the local Chrome browser on the desktop (“Chrome Auto Browse”) and initiate tasks such as apartment viewings or flight bookings.
- Access will be enabled for Google AI Pro users in over 160 additional countries.
- According to Google’s support documentation, the European Economic Area, the United Kingdom, Switzerland, and Nigeria remain excluded—even with the expensive Ultra subscription.
- Spark will redirect you back to confirm payments and other sensitive actions.
Chrome Auto Browse: Spark handles web tasks on its own
Previously, Gemini Spark operated in a remote browser in the cloud. What’s new is that the agent can now use your local Chrome browser on your computer. Once you give your approval, Spark accesses your logged-in accounts and saved passwords and uses the Auto Browse feature that Google introduced for Gemini in Chrome earlier this year. When the agent is active, Chrome displays a corresponding notification at the top of the toolbar.
Google cites typical “tedious web errands” as examples: scheduling viewings for saved apartments or researching flight options and starting the booking process. The Chrome integration is rolling out in the U.S. first, with other regions to follow later.
What exactly is Gemini Spark?
Gemini Spark is Google’s personal AI agent that automates complex, multi-step processes in the background. Instead of just waiting for the next question, Spark works through tasks independently—up to 15 at a time. The agent accesses Gmail, Calendar, Docs, Sheets, Slides, and Drive, as well as Google Search, its own cloud browser, and code execution.
Google distinguishes between three components: A task is the goal (“Plan my business trip to London”), a schedule is a trigger (“If my flight is delayed, suggest a new route”), and a skill is a reusable set of instructions on how the agent should complete a task. Spark launched in May 2026, initially with the expensive Ultra subscription; since then, Google has primarily expanded the Workspace features. If you want to learn about the assistant’s voice features, our article on Gemini for macOS is a good place to start.
Availability: Over 160 countries—just not in Europe
Alongside the Chrome integration, Google is expanding Spark access for Google AI Pro subscribers to over 160 additional countries. That sounds like widespread availability, but there’s a clear limitation. The key statement isn’t in the announcement, but in Google’s own support documentation: Spark is available wherever the Gemini apps run—except in the European Economic Area, the United Kingdom, Switzerland, and Nigeria. This explicitly applies to the Ultra subscription as well, with which the agent would otherwise run worldwide.
For Germany, this means that as of the end of July 2026, Gemini Spark simply does not exist here, regardless of the subscription plan. It’s worth noting that as recently as mid-June, Australia, Canada, India, and Japan were also on the exclusion list. Google has since gradually made the service available in these countries. Europe and Nigeria remain on the list. Google has not provided an official reason.
| Feature | Gemini Spark (as of July 30, 2026) |
|---|---|
| New Feature | Chrome Auto Browse on desktop |
| Base Model | Gemini 3.5 |
| Concurrent Tasks | Up to 15 simultaneously |
| Access to the U.S. | Starting with Google AI Pro (about $20/month) |
| Other countries | Only with Google AI Ultra (about $100/month) |
| Excluded | EEA, United Kingdom, Switzerland, Nigeria |
| Language | English only for now |
Security: Prompt Injection, and you remain in control
An agent with access to logged-in accounts raises security concerns. Google relies on multiple protective mechanisms in Spark. The agent returns sensitive actions—such as payments or submitting forms—to you for confirmation (“Take control” mode). Additionally, the browser protects against so-called prompt injection, in which hidden instructions on a webpage are designed to trick the agent into performing unwanted actions.
However, Google itself emphasizes that these protective features are no guarantee against all risks and that active oversight by the user remains the most important safeguard. So anyone granting full access to an AI agent should pay close attention to what it’s doing.
Alternatives for Germany
Anyone in Germany looking for a comparable AI agent today will find what they’re looking for among the competition. ChatGPT Work from OpenAI and Claude Cowork from Anthropic offer similar background agents, start at around $20 per month, and are available in Europe. You can read more about the wave of agent-based AI in our articles on the ASUS Zenni Claw and the latest price cuts at OpenAI.
Conclusion
With Chrome integration, Gemini Spark becomes much more suitable for everyday use: The agent will handle web tasks directly in your own browser instead of in a isolated cloud environment. For German users, this progress remains theoretical for now, as the EEA exclusion still applies. The interest is there—Google is just sitting on it. Until that changes, the agents from OpenAI and Anthropic are the obvious alternative.
Sources: Google Blog, Google Support
