These last years have seen the multiplication of tra c agent-based frameworks (MATSim, SUMO...). If these frame- works are well- tted for the study of normal tra c conditions, it is often complex to adapt them - in particular for non-computer scientists - for more speci c application contexts such as the study of impacts of uncommon events (e.g. car accidents, technological hazards). In this paper, we present a new open-source (GPL) tool, integrated into the GAMA modeling and simulation platform, allowing to easily de ne new microscopic tra c simulations, easily tunable, with a detailed representation of the driver operational behaviors. In particular, it allows to take into account the road infrastructures and tra c signals, the change of lanes of the drivers and their respects of norms. Moreover, the tool allows to run simulations at city level with tens of thou- sands of driver agents. We illustrate the use of this plug-in through an example for the tra c simulation of the Rouen city (France).