Google's Circle to Search now identifies multiple objects at once, powered by Gemini 3

Circle to Search Gets Major Upgrade with Multi-Object Recognition

Google launched a significant update to Circle to Search, timed with Samsung's Galaxy S26 release, that fundamentally changes how users interact with visual content on their phones. The feature, now available on over 580 million Android devices 3

3

, can identify multiple objects in an image simultaneously thanks to Gemini 3's advanced capabilities. Previously limited to searching a single object, Circle to Search now scans entire scenes at once, whether that's a complete outfit on social media or a complex coral reef photograph 1

1

.

The tool works by tapping and holding your device's home button, then circling the passage or image you want to explore. What sets this Google Search update apart is its ability to handle multi-step reasoning through relationships between objects 4

4

. If you circle a photo of a coral reef and ask how the fish coexist, the AI identifies each species—from Honeycomb Filefish to Moon Jellyfish—and explains the science behind their underwater community 3

3

.

Advanced Shopping Queries Transform the Fashion Experience

The update delivers substantial improvements for shopping experience scenarios. When you spot a creator's outfit on Instagram or any social feed, you can now circle the entire person to deconstruct the look completely 2

2

. Circle to Search identifies each piece—clothing items, shoes, and accessories—providing a comprehensive breakdown of similar products. A new "Find the look" button appears above the redesigned "Ask anything" bar, which has moved from the top to the bottom of the screen, creating a more chat-like user interface 3

3

.

Google integrated its virtual try-on dressing room directly into the search overlay 1

1

. After circling an outfit and tapping "Find the look," users can select a new "Try it on" button to enter a virtual dressing room and test how clothes fit their body type 2

2

. This integration makes impulse buys more tempting by removing friction between discovery and visualization.

Gemini 3 Powers Multi-Step AI Processing

Behind the scenes, Gemini 3's agentic planning, reasoning, and tool calling capabilities drive this multi-object search functionality 3

3

. The system no longer simply looks for a single match. Instead, the AI thinks through a multi-step plan, automatically identifying the most important parts of an image to crop, running several searches at once, and cross-referencing findings to compile a final response 3

3

. This approach enables the tool to handle complex queries across fashion, home decor, and educational contexts.

Google notes this creates new opportunities for merchants and businesses to be discovered through visual search 3

3

. Whether users are curating mood boards for room redesigns or satisfying curiosity about everything in a photograph, the image recognition system processes it all simultaneously.

Galaxy S26 and Pixel 10 Get First Access

These features rolled out on February 25 for Samsung Galaxy S26 series and Pixel 10 devices 2

2

. Google plans Android device integration across more phones soon, though no specific timeline was provided 1

1

. The staggered rollout prioritizes flagship devices first, likely to showcase the processing power required for Gemini 3's complex operations. Watch for expansion announcements as Google extends these capabilities to mid-range Android phones in coming months.