Google is testing a floating search bar for Chrome on Windows, summoned by the Ctrl+Shift+Sp shortcut, that could evolve into an AI-powered desktop assistant. Dubbed “Everywhere Omnibox” and tied to the internal Project Loom, the feature has appeared in Chrome Canary, the experimental channel where Google trials bleeding-edge ideas. The interface, a standalone text box, overlays other applications, letting users search the web or potentially trigger AI actions without switching to a browser window.

The discovery aligns with Google’s aggressive push to embed AI across its products. While the company has not officially announced Everywhere Omnibox, early testers note it resembles a lightweight version of the Chrome address bar—the omnibox—but freed from the browser frame. Its floating nature suggests a reimagined Windows Search or a direct challenger to Microsoft’s Copilot, which also lives outside the browser.

The Shortcut That Unlocks It

In Chrome Canary for Windows, pressing Ctrl+Shift+Sp activates a small overlay near the center of the screen. The field accepts text immediately, and hitting Enter runs a Google search in the default browser. Users can dismiss it with Esc. The current build shows no advanced AI integration—no chat, no contextual responses—but the Project Loom codename hints at something more ambitious. Google registered a trademark for “Loom” years ago, and internal sources have linked it to AI-driven features that stitch together services like Gmail, Drive, and Calendar.

The minimalist interface echoes the omnisearch concept: a single bar that retrieves files, launches apps, and fetches web results. On Linux, Chrome OS’s launcher behaves similarly, and macOS has Spotlight. Windows 11’s own taskbar search has slowly improved, but Google’s entry could offer tighter integration with Google’s ecosystem—similar to how the Chrome omnibox already handles calculations, unit conversions, and site-specific commands.

From Omnibox to Everywhere

Chrome’s omnibox has long been more than a URL bar. It parses natural language, suggests answers from the user’s history, and even performs quick actions. Everywhere Omnibox takes that functionality system-wide. Imagine highlighting text in Word, invoking the floating bar, and asking it to “summarize this” or “translate to French”—all without opening Chrome. While those AI talents aren’t present yet, the groundwork in Chrome Canary suggests they’re coming.

Google has experimented with similar ideas before. The short-lived Google Desktop, launched in 2004, indexed local files and emails for fast retrieval. It was killed after Windows Vista and 7 baked search into the OS. More recently, Chrome Actions let users type commands like “delete history” directly into the bar. Everywhere Omnibox expands the canvas: a persistent, hotkey-driven assistant that can tap into Google’s evolving AI models.

Project Loom and the AI Future

Project Loom is the reported internal codename. The name appeared in Chromium code repositories years ago, initially tied to experimental voice recognition. Now it seems to have broadened into a platform for ambient computing—where AI follows you across devices and apps. Google’s recent demos of Gemini AI, with its multimodal understanding and ability to reason over user data, would be a natural fit. A floating search bar could become the front door to Gemini, answering complex queries, generating images, or even taking actions like booking a flight.

Such integration would mirror what Microsoft has done with Copilot in Windows. Microsoft’s assistant also lives in a sidebar, can be summoned with a shortcut, and can control some system settings. However, Google’s strengths in web search and its massive user base could make a cross-platform, Chrome-driven assistant far more ubiquitous. Unlike Copilot, which is exclusive to Windows 11, Chrome runs on billions of devices, making Everywhere Omnibox a potential Trojan horse for Google’s AI services on desktops.

Practical Implications for Windows Users

For Windows enthusiasts, the feature raises immediate questions. Will it consume resources? Early Canary builds show a lightweight process that uses minimal RAM, but adding AI could change that. Privacy is another concern: a always-listening or always-available search overlay might transmit keystrokes for predictive suggestions. Google will need to be transparent about what data leaves the device.

There’s also the matter of keyboard shortcut conflicts. Ctrl+Shift+Sp isn’t commonly reserved by Windows or major applications, but power users who customise hotkeys may need to remap it. Chrome typically allows such adjustments in advanced settings.

If Everywhere Omnibox graduates to stable Chrome, it could alter workflows dramatically. Developers, writers, and researchers often jump between apps to search. Having a persistent, AI-imbued bar that understands context—like the current document title or selected text—would reduce friction. Microsoft’s own PowerToys Run already offers a similar launcher, but it lacks web search depth and AI reasoning.

Competition in the Desktop AI Space

The desktop overlay is becoming a crowded field. Apple’s Spotlight serves as a system-wide search on macOS, while Raycast and Alfred add extensibility and AI plugins. On Windows, Flow Launcher, Keypirinha, and the aforementioned PowerToys Run offer quick search and app launching. Google’s advantage is its data—a logged-in Chrome user has their bookmarks, browsing history, passwords, and possibly Drive files in reach. When augmented by AI, Everywhere Omnibox could proactively surface that information.

Microsoft, however, is not standing still. Copilot is integrated into Bing, Edge, and the Windows taskbar. It can already generate text, answer questions, and perform limited system commands. The battle will likely center on which ecosystem users live in: Microsoft’s Graph-connected productivity suite or Google’s web-centric services.

What to Expect in Future Canary Builds

Chrome Canary updates daily, so the feature could evolve quickly. Observers should watch for toggles that enable AI backends, voice input, or system-level permissions that let the overlay interact with files and windows. Google often uses field trials to test features with a subset of users, so a controlled rollout might happen before any public announcement.

Developers who dig into Chromium commits may find clues. Past Project Loom traces appeared in code related to “FloatingWindow” and “OmniboxOverlay.” The Ctrl+Shift+Sp shortcut was added in a recent commit that explicitly mentions Windows as the target platform, suggesting a more polished implementation than a simple prototype.

The Road Ahead: Possible Release Timeline

Google rarely commits to launch dates for Canary experiments. Some features, like tab groups or memory saver, spent months in Canary before reaching stable. Others vanish without a trace. Given the competitive pressure from Microsoft and OpenAI, however, Google might accelerate an AI-infused desktop search tool. The timing also coincides with the rollout of Chrome’s built-in AI features like writing help and tab organizers, which use on-device models.

A reasonable guess: an alpha feature flag in Chrome 130 or later, with a wider beta in early 2025, and a stable launch by mid-2025. That would align with Google’s typical cadence and the anticipated next wave of Gemini upgrades.

Privacy and Security Considerations

A floating overlay that can read active window titles or selected text would require sensitive permissions. Chromium’s security model normally sandboxes web pages, but a system-level assistant needs deeper access. Google might implement an opt-in permission dialog and restrict data processing to on-device modes where possible. The company has emphasized its commitment to privacy in AI features, such as the on-device translation in Chrome, and the same approach could apply here.

Still, skeptics will worry about Google harvesting more user interaction data. Transparency logs and clear documentation around which keystrokes are sent to servers will be critical for trust. Enterprise administrators, in particular, will demand group policies to disable the feature if it ships.

How to Try Everywhere Omnibox Right Now

As of this writing, the feature can be enabled by installing Chrome Canary and navigating to chrome://flags. Search for “Everywhere Omnibox” or “Project Loom”—the flag name may differ by build. Enable it, restart Chrome, and press Ctrl+Shift+Sp anywhere on the Windows desktop. Note that background Chrome processes must be running; the feature doesn’t work if Chrome is fully closed. Early adopters report occasional glitches, such as the bar appearing behind some full-screen apps, so it’s clearly a work in progress.

The Bigger Picture: Chrome as a Platform

This experiment underscores Chrome’s transformation from a simple browser into a platform. Google has steadily added OS-like capabilities: progressive web apps, file system access, and background sync. An always-available search overlay blurs the line between browser and desktop, much like Chrome OS does on Chromebooks. On Windows, it could position Chrome as the shell for a Google-centric experience, bypassing Microsoft’s own UI.

For Windows users, the prospect is exciting but also raises questions about vendor lock-in. If the tool ties deeply into Google services, moving to another browser could feel like losing a key productivity layer. That said, competition often spurs innovation, and Microsoft’s Copilot may become even more capable as a result.

Conclusion

Google’s Everywhere Omnibox in Chrome Canary is a clear signal that the search giant wants a larger footprint on the Windows desktop. The Ctrl+Shift+Sp shortcut, the floating design, and the ties to Project Loom suggest a future where AI-assisted search is always a keystroke away—regardless of which app you’re using. While it’s still a barebones prototype, the potential to integrate Gemini and other Google services could reshape how we interact with our PCs. Windows enthusiasts should keep a close eye on Canary builds, as the next few months will likely reveal whether this feature becomes a core part of Chrome or fades as another experiment.