Google is quietly building a feature that could fundamentally change how people interact with AI on their phones. Rather than typing out elaborate descriptions of what they want Gemini to analyze in an image, users will soon be able to simply draw on it.
The feature, spotted in a teardown of the latest Google app beta, lets users annotate images with freehand markup before sending them to Gemini. Circle a plant you can’t identify. Underline a line of code that’s throwing errors. Draw an arrow pointing to the crack in your foundation you want assessed. It’s the kind of natural, intuitive interaction that bridges the gap between how humans think about visual problems and how they’ve historically had to communicate them to AI.
According to Android Authority, which first uncovered the feature through an APK teardown of Google app version 16.26.33 beta, the markup tool includes a pen for freehand drawing and a text tool for adding typed annotations directly onto images. The interface appears polished enough to suggest it’s fairly far along in development, though Google hasn’t officially announced it.
This isn’t the first time evidence of image markup has surfaced. Android Authority noted that traces of the feature appeared in an earlier teardown back in March 2025, but the latest version shows significant refinement. The tool now seems to be integrated into Gemini’s image-sharing flow on Android, appearing as an option when users attach photos to their prompts.
The logic behind it is straightforward. AI image analysis has gotten remarkably good, but it still struggles with ambiguity. Tell Gemini “what’s wrong with this picture” and it might identify five things you didn’t care about while missing the one thing you did. Markup solves that. It gives users a way to direct the AI’s attention with precision — no verbose prompt engineering required.
Why Visual Annotation Matters More Than It Seems
Google has been on an aggressive campaign to make Gemini the default AI assistant across Android, and this markup tool fits squarely into that strategy. The company has spent the past year weaving Gemini into virtually every surface of the Android experience — from the home screen overlay to Google Messages to the camera interface. Each integration is designed to reduce friction between a user’s intent and the AI’s response.
The markup feature attacks one of the last remaining high-friction interactions: visual queries. Google Lens already lets users point their cameras at objects for identification, but Lens operates more like a search engine than a conversational partner. Gemini’s markup tool, by contrast, keeps users inside a dialogue. You annotate, you ask, you get an answer, you follow up. The conversational thread stays intact.
This matters for professional use cases especially. Contractors could photograph a job site and circle specific problem areas for Gemini to assess. Designers could mark up screenshots with annotations asking the AI to critique specific layout choices. Medical professionals — within the bounds of appropriate use — could highlight regions of concern in clinical images. The applications multiply quickly once you give users a pointing mechanism that’s more expressive than a tap and more efficient than a paragraph of text.
And Google isn’t operating in a vacuum here. Apple has been steadily expanding its own AI image capabilities through Apple Intelligence, and OpenAI’s ChatGPT already supports image uploads with conversational analysis. But neither currently offers a native markup layer integrated directly into the prompt interface. If Google ships this feature broadly, it would represent a genuine differentiator — at least temporarily.
The timing is notable. Google I/O 2025 showcased a barrage of Gemini upgrades, including expanded multimodal capabilities and deeper Android integration. The markup tool wasn’t among the announcements, but its appearance in beta code shortly after the conference suggests it could be part of a near-term rollout. Google often seeds features into beta builds weeks before public launch.
There’s also a competitive dimension worth watching. Samsung and other Android OEMs have been building their own AI features that sometimes overlap with — or outright duplicate — Google’s Gemini capabilities. Samsung’s Galaxy AI, for instance, includes a “Circle to Search” feature that lets users draw around on-screen elements to trigger a Google search. Gemini’s markup tool takes that concept further by enabling not just search but full conversational analysis of annotated images.
So where does this leave users? Waiting, mostly. The feature hasn’t been officially confirmed by Google, and APK teardowns don’t guarantee a public release. Features get shelved all the time. But the level of polish Android Authority observed in this latest teardown — complete with functional pen and text tools — suggests Google is serious about shipping it.
For Android users who already rely on Gemini for image-related queries, the markup tool promises to eliminate one of the assistant’s most persistent annoyances: the need to over-explain what you’re looking at. A circle and a question mark might soon be all it takes.
That’s a small change in interface design. But it could meaningfully alter how millions of people use AI every day.
WebProNews is an iEntry Publication