In this demo paper we present MultiCAMBA (Multi-CAMera Broadcasting Assistant), a context– and location–aware system that, using a 3D model updated continuously with real-time data retrieved from the scenario, helps technical directors (TDs) in the live broadcasting task. They can indicate in run-time their interest in certain moving objects or geographic areas, and the system is in charge of selecting the cameras that can provide the kind of view required. To achieve this task, the system continuously recreates the views of the cameras in a 3D scenario, considering possible occlusions among the objects.