A detailed After Action Review (AAR) is essential to get the most out of a live exercise or training simulation. An effective AAR requires an overall picture of all the actions and events that took place during the exercise. Compiling such an overall picture is often a complicated task, because the different participants all perform their own data logging, resulting in many separate data logs, often in different data formats. TNO’s Generic Reconstruction And Computing Environment (GRACE) has been developed for quickly compiling the overall picture of a live exercise or training simulation. GRACE can process various data formats and provides various tools for the analysis, integration, and visualization of logged data, making it possible to create a complete reconstruction of an exercise. Its flexible architecture allows the addition of new data formats or new analysis tools. GRACE has been proven successful during several live firings. This paper describes the innovative architecture and ideas behind GRACE. We will discuss how GRACE handles, interprets and combines data, how the framework was designed to be generic and flexible for extensions, and we will give examples of our experiences with the tool.