One of the consequence of the scaling down of latest technologies, is that digital circuits are more prone to be affected by faults caused by physical manufacturing defects, environmental perturbations (e.g., radiations, electromagnetic interference), or aging-related phenomena. Understanding the behavior of the whole system in the presence of faults affecting digital circuits is crucial for designing and validating fault tolerance techniques. Fault injection is a well-known and widely used approach for estimating the behavior of a system in the presence of errors. Among the different techniques, FPGA-based fault injection is one of the most popular since it allows a low-cost, rapid and accurate approach compared to simulation- and hardware based fault injectors. In this paper we present a flexible FPGA fault injector able to inject in the memory hierarchy of a given system.