Google launches Guided Vision in Gemini Live, giving real-time audio descriptions through the Android camera

Google is launching Guided Vision in Gemini Live on compatible Android devices, a feature that gives real-time audio descriptions of anything a user points their phone camera at. The rollout began October 1, 2026, according to The Verge.

What Guided Vision does

Users share their camera in Gemini Live and the assistant describes what it sees. Google lists these use cases:

• Reading small text, such as fine print
• Describing the user's surroundings
• Finding or identifying nearby objects
• Describing details on specific objects

Users can ask follow-up questions about what is in view. One example is asking Gemini to read the expiration date on an item it helped locate. Google says Gemini gives audio cues to help line up the camera when a user asks about something that is not in the frame.

Where it is available

Guided Vision is available in three places:

1. The Gemini app, through Gemini Live
2. Google TalkBack, Android's screen reader
3. An accessibility shortcut that can be set up in the Settings app on devices running Android 9 and above

The feature was announced as part of a recent set of Pixel updates. Google has not specified which devices count as "compatible" for the Gemini Live integration. The Android 9 and above requirement applies to the Settings shortcut.

Limits Google states

Google cautions users not to rely on Guided Vision for "navigation, safe-travel guidance, or obstacle detection." It also says the feature is not a replacement for a cane or other mobility aid.

What was not disclosed

The source material does not identify the Gemini model version behind Guided Vision. It also gives no latency figures, accuracy benchmarks, supported languages, or regional availability. No separate pricing has been disclosed, and Google has not said whether the feature requires a paid Gemini plan.

Competitive context

The Verge compares Guided Vision to VoiceOver Live Recognition, which Apple has added to the iPhone and Vision Pro. Both features target people who are blind or have low vision, as well as users who want help in specific situations.

What this means

This is a product update, not a model release. Google has not announced new weights, a new model version, or new API capabilities. The feature packages Gemini Live's existing camera-sharing capability into an accessibility workflow.

The distribution matters more than the capability. Putting the feature in TalkBack and a system-level Settings shortcut makes it reachable without opening the Gemini app, which is likely to matter for users who depend on screen readers. It also puts Google and Apple in direct competition on multimodal accessibility features built into the operating system.

The explicit warning against using it for navigation or obstacle detection is the key limit. It suggests Google does not consider the system reliable enough for safety-critical use. Without published accuracy or latency data, users and developers have no way to measure how reliable the text reading and object identification are. Independent testing of text-reading accuracy on small or low-contrast print would be the most useful next step.