According to The Verge, Google has begun rolling out real-time screen and camera analysis features for Gemini Live, available now to select Google One AI Premium subscribers. The update, powered by last year’s Project Astra, enables the AI to interpret visual inputs and answer questions on the fly, a company spokesperson confirmed.
A Reddit user demonstrated Gemini’s screen-reading capability on a Xiaomi device this week, sharing a video of the AI identifying app icons and summarizing content. 9to5Google first spotted the post, which aligns with Google’s March announcement that these features would launch by month’s end.
In a company-published example, a user points their phone camera at a pottery piece and asks Gemini to suggest paint colors. The AI analyzes the glaze texture and surface area before recommending options.
The rollout gives Google an edge as rivals scramble: Amazon’s Alexa Plus enters limited testing soon, Apple delayed its Siri overhaul, and Samsung—despite maintaining Bixby—defaults to Gemini on its devices. All aim to match Astra’s real-time visual processing, which Google deployed nearly a year after its initial demo.
Featured image credit: Google