Google's Circle to Search can now hunt down multiple items at once

If you use Circle to Search regularly, you have probably run into this headache: you see a great outfit on Instagram and want to find the sneakers, but your circle accidentally grabs the hem of the coat, confusing the algorithm entirely.

Alongside Samsung’s Galaxy S26 launch, Google rolled out a major upgrade to Circle to Search that lets it search multiple objects in a single image simultaneously.

Deconstructing full outfits in one circle

The feature works just like before, but what happens under the hood is vastly different.

Instead of carefully cropping a single item, you can now scribble over an entire look. Circle to Search will automatically deconstruct the outfit, identifying the outerwear, bottoms, shoes, and accessories in parallel, and serve up matching product results for every piece.

In regions where Google Shopping already supports virtual try-on, users can even tap Try On right from the search overlay to preview the clothes on their own digital avatar.

It is not just for fashion either. Google showed a demo where someone circles a crowded aquarium shot and asks what the fish are and how they coexist. The tool immediately identifies individual species, like the Moon Jellyfish and Honeycomb Filefish, while providing an AI-generated breakdown of their ecosystem alongside web links.

Powered by Gemini 3's visual agent

Google says this multi-object leap is powered by Gemini 3's agentic planning and reasoning, paired with a visual query fan-out system.

Essentially, the model breaks down the task like a researcher. It scans the selected area, crops the most relevant subjects automatically, fires off multiple parallel web searches, and cross-references the findings before assembling a single structured response.

Multi-object search is rolling out first on the new Samsung Galaxy S26 series and Google Pixel 10 devices, with wider Android availability coming later. Alongside it, Google announced a background agent beta for ridesharing and grocery runs, plus on-device call scam detection for Samsung phones.

Why it matters

For consumers and visual curators, this completely eliminates the tedious dance of zooming, cropping, and searching four times just to save a mood board. It is immediately useful, frictionless, and cuts visual shopping friction in half.

For the retail and advertising industry, it opens up massive inventory discovery. Secondary items like belts, watches, or layering pieces that used to get ignored in visual searches now share the screen with hero products, which will likely force merchants to rethink product placement in digital campaigns.

We think Google played this smartly by embedding agentic workflows into an existing, widely adopted gesture instead of pushing another separate chatbot app. The main risk looks to be latency and consistency across lower-end devices: firing multiple visual queries at once is computationally heavy, and if the experience lags on older hardware, users will simply revert to manual searches.