Live
Built withYOLOXONNX RuntimeOpenCVWebSockets
Kestrel
Crowd and traffic analytics that runs without a GPU
LeafWise
Crop disease detection that works with no signal
leafwise-scan.vercel.app
- Model size, running fully in-browser
- 9.25 MB
- Across 14 crop types
- 38 classes
Overview
What LeafWise does.
A farmer photographs a leaf and gets an answer without a network round trip, because the model runs in the browser. A 9.25 MB model covers 38 disease classes across 14 crops, and accuracy is published against field photographs as well as the training set — because a single number would flatter it.
Runs entirely in the browser, offline after first load
38 disease classes across 14 crops
Cross-dataset accuracy published, not just in-sample
No image ever leaves the device
Keep exploring
More in Vision & Edge AI
LiveBuilt withMediaPipeTypeScriptWebGLReactKinetix
Movement analysis in the browser, on-device
Let's talk
Have a problem worth solving?
Tell us what you're building. We'll help you scope it, build it, and ship it.