Google is rolling out Guided Vision, a new Gemini Live feature that turns your phone’s camera into a real-time audio guide. The Verge AI reports that it starts launching today on compatible Android devices. You point your camera at something and Gemini describes what it sees, reads small text and helps you find objects nearby.
Google built it mainly for people who are blind or have low vision. It’s also useful for anyone who just needs help in the moment, like when the print on a medicine bottle is too small to read.
5 Things to Know About Guided Vision
It reads, describes and finds things in real time
Once you share your camera in Gemini Live, Google’s AI can help with four main jobs:
- Reading small text
- Describing your surroundings
- Finding or identifying objects around you
- Describing details on specific objects
That’s a broad set of everyday tasks. Labels, menus, mail and packaging all fall into the “I can’t quite make this out” category, and that’s where an AI narrator saves the most time.
You can keep asking follow-up questions
Guided Vision isn’t a one-shot description. You can keep the conversation going about whatever’s in front of the camera. The Verge AI gives a practical example: once Gemini helps you find an item, you can ask it to read off the expiration date.
This is the part that matters most. Older image description tools give you one caption and stop. A back-and-forth conversation is much closer to how you’d ask a sighted friend for help.
It helps you aim the camera
Aiming is a real problem for blind users. If you can’t see the screen, how do you know the object is in frame? Google says Gemini will give audio cues to help you line up the camera when you ask about something that isn’t in the shot.
It sounds like a small detail, but it’s the difference between a feature that works in a demo and one that works on an actual kitchen counter.
You can open it from several places
Google is making Guided Vision easy to reach. You can open it through:
- The Gemini app, via Gemini Live
- Google TalkBack, Android’s built-in screen reader
- An accessibility shortcut in the Settings app, on devices running Android 9 and above
Google first announced the feature in a recent batch of Pixel updates, according to The Verge AI. Support going back to Android 9 is a smart choice. A lot of people who rely on accessibility tools aren’t using the newest flagship phones.
Google is clear about what it shouldn’t do
There’s a big caveat here. Google warns users not to rely on Guided Vision for “navigation, safe-travel guidance, or obstacle detection.” It also says the feature isn’t a replacement for a cane or other mobility aid.
That’s the responsible call. AI vision models can still misread a scene or miss something, and the stakes are very different when someone’s crossing a street versus checking a cereal box. For now, Guided Vision is a reading and identifying tool, not a safety tool.
How It Stacks Up Against Apple
Google isn’t first here. Apple has already added VoiceOver Live Recognition to the iPhone and Vision Pro, and it serves a similar purpose for blind and low-vision users. The Verge AI draws that direct comparison.
What stands out with Google’s version is that it runs on Gemini’s conversational model. You aren’t stuck with fixed recognition categories. You can just ask whatever you want to know. Apple’s Vision Pro does have a headset form factor, though, which means hands-free use, and phones can’t match that.
Why This Matters
Accessibility is quietly turning into one of the best real-world uses of multimodal AI. Live camera assistants have mostly been pitched as flashy demos, like identifying landmarks or explaining a math problem. Guided Vision points the same technology at a group of people who get daily, practical value from it.
It also shows a trend. Big platforms are building AI features straight into the operating system and its accessibility settings instead of hiding them in a separate app. That cuts friction for the people who need these tools most.
The next question is reliability. How well Guided Vision handles bad lighting, cluttered rooms and odd fonts will decide whether people use it every day or just try it once. Check out The Verge AI for the full details on the rollout.