Google Gemini could be about to take this idea right out of Circle to Search’s playbook

Give Gemini an image, and it’s already quite good at figuring out what it’s looking at. But what if you don’t want it looking at everything in the image, or at least not looking at it all equally? If you can imagine how it might be useful to direct Gemini’s attention to one particular part of a picture, we’ve got just the tool for you, as we spot Google’s work to add markup tools to Gemini image input.

As you’re probably well aware, it’s quite easy to upload an image to Gemini and ask questions about it. But looking through the changes in version 16.42.61.sa.arm64 of the Google app, we’ve been able to identify a slight tweak to that workflow that would let you draw on an image after selecting it from your gallery or shooting a new photo with your camera.

Right now, this feels like a bit of a generic interface rather than one intentionally for communicating with Gemini. For instance, it’s not yet clear why we have the ability to choose different colors here.

Once you do highlight a portion of your image, you can start using that for your Gemini queries:

Even without more explicit directions, Google’s AI is smart enough to know that the part with the circle around it is something you’re trying to draw attention to. But you’re also free to directly tell it to perform an operation based on a highlighted region, as you can see in this example:

Here we’re tapping into Nano Banana’s image editing capabilities and using this input flow to quickly remove unwanted content from a screenshot.

As you can see, this has the potential to be a pretty useful addition to Gemini’s toolset. We wouldn’t be surprised if Google further modifies how you access it and mark your images up in the first place, but hopefully we’re able to see something like this go live in the not-distant future.