This is an exciting time for tuberculosis (TB) diagnostics. The technology for rapid diagnosis of TB and rifampicin (RMP) resistance in pulmonary sputum smear-positive specimens is well advanced, and assays have high specificity with good sensitivity. Nevertheless, the current sensitivity of TB detection means that these assays still cannot replace the standard diagnostic methods for TB or conventional drug susceptibility testing (DST). In extra-pulmonary specimens, the performance of molecular tools varies and should be considered separately for each specimen type. Evidence for the use of these assays for TB and drug resistance detection in individuals co-infected with TB and the human immunodeficiency virus (HIV) is limited. As the positive predictive value for RMP resistance reaches ≥ 90% only when the prevalence of RMP resistance in new TB patients is >15%, which is rare globally, many cases with such resistance will be false-resistant, emphasising the need for a secondary confirmative test. Similarly, increased (or incorrect) diagnosis of TB may compromise programme effectiveness by increasing the numbers of individuals requiring anti-tuberculosis treatment, unless it is carefully planned. For the future, 1) assays with greater sensitivity for TB detection are needed; 2) rapid diagnostics for paediatric TB are important, and there is a need for carefully designed studies, including those involving HIV-positive children; 3) more clinical data need to be obtained from longitudinal studies, especially related to the influence of rapid diagnostics on disease outcome; and 4) point-of-care tests using untreated sputum, blood or urine and little or no equipment would be of immeasurable benefit. Although great progress has been made, we are not there yet.