Chrome’s Address Bar Gets a Permanent AI Sidekick in Latest Test Build

Google is quietly turning its browser into a round-the-clock AI assistant. The latest Chrome Canary build includes a hidden flag that plants a dedicated AI Mode button right inside the address bar. Flip the switch, and a new icon appears. Click it, and you’re talking to Gemini without ever leaving the page you’re on.

This isn’t just a simple shortcut. Testers have spotted floating search windows, the ability to pull context from open tabs, support for images and files, and even early signs of agentic behavior that can act across multiple websites. Chrome, already the world’s most popular browser, is becoming a laboratory for how AI will reshape everyday computing.

The flag, first reported by Android Authority, lives under chrome://flags as “AI Mode omnibox entry point.” Once enabled and the browser restarted, a sparkly icon sits next to the address bar. Clicking it opens a sidebar or dedicated pane where users type prompts backed by Google’s latest models.

Google’s own documentation describes the vision: “With advanced multimodal capabilities, AI Mode can deeply understand your question, no matter how complex it is.” The system can reference open tabs, uploaded images, and files. Ask about a recipe while several cooking sites are open, and the AI pulls details across them without forcing you to copy and paste.

Not everyone is thrilled. The icon sticks around persistently, and some users report hunting through dozens of flags to disable it. Privacy questions linger: when AI Mode scans your open tabs, what data travels back to Google’s servers? The company has added settings under Appearance and AI Innovations, but even those don’t erase every trace.

Meanwhile, the experiments keep coming. A floating desktop AI search bar triggered by Ctrl+Shift+Space on Windows puts AI Mode front and center, pushing traditional search results to the background. Early footage shows users issuing complex prompts that spawn follow-up actions without leaving the floating interface.

Behind the scenes, lighter on-device models like Gemini Nano handle faster responses, though they require hefty downloads. Multimodal features let users feed screenshots, PDFs, or live tab content into prompts. And agentic abilities, currently limited to paid tiers, allow the system to click links, fill forms, and complete tasks with minimal guidance.

Chrome’s transformation reflects a broader shift. The browser no longer just transports users between sites. It interprets, summarizes, and acts. The address bar evolves from a destination field into a prompt box. Tabs become context for smarter answers. The line between search, browsing, and conversation continues to blur.