It's thought to be a component that fits into existing applications. In each platform it supports, OmniGUI will render inside a UIView, View, Control... or whatever a "UI component" is called in the target platform.

It will take the client area assigned to it and render like any other component, using drawing primitives available in the platform.

The interaction with it is made possible because there are abstraction layers that provide the signals OmniGUI needs to make its components act in consequence to user actions.

For drawing, for instance, it uses CoreGraphics in iOS, in Android it uses Canvas, in UWP it uses Win2D and in WPF it uses a DrawingContext.