We present ESO-MAPF, a research and educational platform for experimenting with multi-agent path finding (MAPF). ESO-MAPF focuses on demonstrating the planning-acting chain in the MAPF domain. MAPF is the task of finding collision free paths for agents from their starting positions to given individual goals. The standard MAPF uses the abstraction where agents move in an undirected graph via traversing its edges in discrete steps. The discrete abstraction simplifies the planning phase however resulting discrete plans often need to be executed in the real continuous environment. ESO-MAPF shows how to bridge discrete planning and the acting phase in which the resulting plans are executed on physical robots. We simulate centralized plans on a group of OZOBOT Evo robots using their reflex functionalities and outputs on the surface of the screen that serves as the environment. Various problems arising along the planning-acting chain are illustrated to emphasize the educational point of view.