The integration of context-aware applications with spatial context models is often done using a common query language. However, algorithms that estimate and reason about spatial context information can benefit from a tighter integration. An object-oriented API makes such integration possible and can help reduce the complexity of algorithms making them easier to maintain and develop. This paper propose an object-oriented API for context models of the physical environment and extensions to a location modeling approach called geometric space trees for it to provide adequate support for location modeling. The utility of the API is evaluated in several real-world cases from an indoor location system, and spans several types of spatial reasoning algorithms.