You can access the core interface through a simple C-API, which allows you to start, modify, or stop a tracking configuration.
The pure C-API is by purpose kept low level and designed to enable universal interfacing. The main drawback is that the resulting code appears limited, hard to read, and not easy to maintain. Note, that all other interfaces are built on top of this API.
Tracking states, error messages and the image of the camera can be accessed by registering callback functions. If you use the C-API you have to take care of registering and unregistering your listeners and object pointers.
Here you can find some getting-started articles: