Google's Chrome security team has described a test environment called the Mirror Dimension, a synthetic version of the web for evaluating browser-using AI agents. The simulated sites are designed to resemble real services closely enough for agents to practice multi-step tasks. Google says network requests are intercepted and routed to generated pages instead of being sent to public websites.
That setup lets engineers examine how an agent behaves around realistic pages without letting a test purchase, download, or message affect a real account. Google presents the system as an evaluation tool, not a new browser feature for consumers. The company says the challenge is making the simulated pages convincing enough that an agent behaves naturally while keeping every action inside the test boundary.
Why it matters
Browser agents need testing that includes ordinary web friction, such as navigation and forms, but testing against live services can create real side effects. A simulated web can reduce that risk and make repeated evaluations easier. It still cannot reproduce every behavior or security issue of a changing public website, so results from the sandbox need to be paired with other forms of testing.
Sources and further reading
This story is written in its own words. These links identify the reporting and announcements used to verify the details above.




