WebAI StudioUnderstand the context, joke, or cultural reference in a meme.
On-device vision models can quickly analyze complex visual humor without sharing your browsing habits.
This demo requires an On-Device Language Model with vision capabilities. Please ensure you are using a supported browser with the correct experimental flags enabled.
or click to browse
Upload an image and I'll break down the humor or reference for you.
56 demos, all running on-device.
Drop in your photos, and search them by meaning — "food on a table", "someone smiling" — without one pixel leaving your device.
Point your camera at anything, freeze a frame, and ask the on-device model about what it sees.
You sketch, Gemini Nano guesses: the model interprets your doodle after every stroke.