Vehicle sharing systems, such as bike, car or motorcycle sharing systems, are becoming increasingly popular in big cities as they provide a cheap and green means of mobility. The effectiveness and efficiency, and thus, the quality of service of such systems depends, among other factors, on different strategic and operational management decisions and policies, like the dimension of the fleet or the distribution of vehicles. In particular, the problem of agglutination of available vehicles in certain areas whereas no vehicles are available in other areas is a common problem that needs to be tackled by the operators of such sharing systems. Recently, research works have been focusing on adaptive strategies to reduce such imbalances, mainly through dynamic pricing policies. However, there is no best operational management strategy for all types of bike sharing systems, so it is of foremost importance to be able to anticipate and evaluate the potential effects of such operational management strategies before they can be successfully deployed in the wild. In this paper we present Bike3S, a simulator for a stationbased bike sharing system. The simulator performs semi-realistic simulations of the operation of a bike sharing system in a given area of interest and allows for evaluating and testing different management decisions and strategies. In particular, the simulator has been designed to test different station capacities, station distributions, and balancing strategies. The simulator carries out microscopic agentbased simulations, where users of different types can be defined that act according to their individual goals and objectives which influences the overall dynamics of the whole system.