Getting started
Open the app normally or say “Hey Google, open Vision Lab Assist.” Choose Open camera, allow camera access, point the phone toward the area you want to understand, then capture. Analysis and spoken feedback begin automatically.
ACCESSIBLE SCENE RECOGNITION
See what your image contains.
Capture or add an image to identify nearby objects, understand their position, hear spoken results, and determine your current location.
Choose an image
Drag and drop, paste, upload, or take a photo.
Scanning image
ABOUT & USAGE
Vision assistance, designed around touch and speech.
Camera and results
Double-tap the camera preview or double-press either volume key to take a photo. The description names detected objects, their approximate position in the frame, recognition confidence, your place name, GPS coordinates, and reported location accuracy.
Touch gestures
Double-tap an empty viewer to open the camera. Double-tap a loaded image or swipe up to analyze it. Swipe right to open the camera. Swipe left to repeat the latest result. Swipe down, or triple-tap the status area at the top, to hear the complete usage guide.
Quick Access
Enable Vision Lab Quick Access in Android Accessibility settings to open the camera from any screen by double-pressing Volume Up. This optional service listens only for that key gesture and does not read screen content.
Mistouch protection
Repeated taps cannot start overlapping captures or analyses. Clearing a photo requires pressing Clear twice within a short confirmation period. Hardware actions require a deliberate double-press.
Permissions and privacy
Camera access is used only while the camera view is open. Precise location is requested when a captured scene is analyzed. Enhanced recognition, place lookup, and natural speech require an internet connection and send the minimum information needed to process those features.
Offline behavior
If enhanced recognition is disabled or unavailable, the bundled detector can identify common object categories on the device. Device speech is used automatically if online speech is unavailable. Place-name lookup still requires a connection.
Important safety note
Results can be incomplete or incorrect, and GPS accuracy varies by device and surroundings. Use Vision Lab Assist for additional awareness, not as a replacement for a cane, guide dog, trained human assistance, or safe navigation judgment.