What is a cure and how do we know it? This essay examines the history of tuberculosis research to trace how ideas of cure shifted from Robert Koch’s development of tuberculin in late nineteenth-century Berlin to the earliest randomized studies of the efficacy of antibiotics in midcentury Britain and India. What sorts of evidence were required to know that a cure was in fact a cure? How did scientists ascertain when and if a cure had failed? And what were the ethical entailments of such modes of knowing? Through this history, I suggest that cure, rather than signaling a definitive end to illness, might instead be figured as an ending lacking finality. Like promises, cures for tuberculosis can be broken. Thinking at the limits of cure makes possible a reflection on the persistence of “curable diseases” into the present, as well as on the ways in which therapeutic knowledge comes to be consolidated and dismantled.