WebAI StudioBack to Demos
Draw & Guess
Image InputYou sketch, Gemini Nano guesses: the model interprets your doodle after every stroke.
Why On-Device?
Guessing on every stroke only works when vision inference is local — no upload per doodle, no latency, no cost per guess.
Interactive Demo
Prompt API (image): Loading...
Draw this:
0 / 0 rounds wonSketch "" here — mouse, finger, or stylus
0 strokes
Model Guesses
Draw a stroke — the model guesses after each one.
Implementation Code
demo.js
Keep exploring
56 demos, all running on-device.
Photo Semantic Search
Drop in your photos, and search them by meaning — "food on a table", "someone smiling" — without one pixel leaving your device.
Prompt API Semantic Embedder
Image Description
Generate detailed alt-text or descriptions for images.
Prompt API
Explain a Meme
Understand the context, joke, or cultural reference in a meme.
Prompt API