The advent of the Xpert(R) MTB/RIF technique was a revolution in the diagnosis of tuberculosis, especially in areas with high incidence and low resources. It allows the detection of Mycobacterium tuberculosis complex and simultaneously the most common resistance mutations to rifampicin in less than 2h. For respiratory samples the sensitivity is very high, but it decreases for extrapulmonary samples and children. Although it is faster and simpler than conventional methods, it presents some limitations and new and better techniques are needed to reduce the number of cases and deaths caused by tuberculosis. This review aims to assess the scientific evidence around the diagnostic performance of Xpert(R) MTB/RIF in different types of samples and populations, as well as analyse its strengths and limitations for TB diagnosis.