Type a prompt, allow the camera, and a live AI vision application starts on this device. Nothing is uploaded. Everything runs here.
What should we build together?
Or teach it to find something specific
Draw a box around the exact thing once: an object, a part, a specific person. Four guided steps train a model on this device that finds it live. Nothing is uploaded.
Detection mode runs an on-device model that recognizes 80 everyday objects. Teach mode trains your own model right here. On the full platform, Lightapp builds production models for your exact products and processes.
Your prompt
Lightapp is building your application
The model runs entirely in your browser. First load is a few megabytes.
Lightapp
Your applications and trained models are saved to your account, so they are waiting for you on any device.
A Lightapp demonstration. Video never leaves your device.