Ļéàŕñ ķàþî › Explore the engine

Vision

Text and layout recognised in an image, on your device.

The native kapi-vision plugin reads text and page layout from images with two ONNX models. This playground runs the same models in the browser: recognition first, layout on request. Pick an image, or drop one of your own; select a box on the image or a line in the list and the other highlights.