Gemini Live's Guided Vision Turns Your Phone Camera Into a Narrator

Google has started rolling out Guided Vision, a new accessibility feature inside Gemini Live that describes the world through your phone's camera in real time. It's available from today on compatible Android devices.

The idea is simple. Open Gemini Live, share your camera, and the AI talks you through what it sees. Google says it can read small text such as the fine print on a label or contract, describe your surroundings, help you find or identify objects nearby, and pick out details on a specific item you point at.

The feature is aimed first at people who are blind or have low vision. Google also positions it for anyone who wants a hand in a particular moment, such as squinting at a tiny ingredient list or hunting for something in a cluttered room.

Guided Vision lands in territory Apple has already entered. The iPhone and Vision Pro both offer VoiceOver Live Recognition, which similarly uses the camera to describe scenes and objects aloud. Google's version leans on Gemini's conversational abilities, so users can in theory ask follow-up questions instead of only listening to a fixed description.

Beyond the Gemini app itself, Google says the feature will also appear in additional places, though the full list of surfaces was not detailed in the source summary.

As with any AI-driven description tool, accuracy will matter a great deal. Misreading a medication label or misidentifying an object carries more weight for users who rely on the feature as a primary source of visual information. Real-world testing from the accessibility community will show how reliable it is outside demos.

For now, Guided Vision is another sign that the big platform companies see live, camera-based AI assistance as a core accessibility tool, not just a novelty feature for general chatbots.