The World Health Organization recommends standardised treatment durations for patients with tuberculosis. We identified and validated a host-RNA signature as a biomarker for individualised therapy durations for patients with drug-susceptible (DS)- and multidrug-resistant (MDR)-tuberculosis.Adult patients with pulmonary tuberculosis were prospectively enrolled into 5 independent cohorts in Germany and Romania. Clinical and microbiological data, and whole-blood for RNA transcriptomic analysis were collected at pre-defined timepoints throughout therapy. Treatment outcomes were ascertained Treatment outcomes were ascertained by TBNET criteria (6-month culture status/one-year follow-up). A whole-blood RNA therapy end model was developed in a multi-step process involving a machine-learning algorithm to identify hypothetical individual end-of-treatment timepoints.Fifty patients with drug-susceptible (DS)-tuberculosis and 30 patients with MDR-tuberculosis were recruited in the German identification cohorts (DS- and MDR-GIC), 28 patients with DS-tuberculosis and 32 patients with MDR-tuberculosis in the German validation cohorts (DS- and MDR-GVC), and 52 patients with MDR-tuberculosis in the Romanian validation cohort (MDR-RVC). A 22-gene RNA model that defined cure-associated end-of-therapy timepoints was derived from the DS- and MDR-GIC data. The model was superior to other published signatures to accurately predict clinical outcomes for patients in the DS-GVC (AUC=0.94 [95%CI:0.9-0.98]) and suggests that cure may be achieved with shorter treatment durations for tuberculosis patients in the MDR-GIC (mean reduction 218.0 days, 34.2%, p<0.001), the MDR-GVC (mean reduction 211.0 days, 32.9%, p<0.001), and the MDR-RVC (mean reduction of 161.0 days, 23.4%, p=0.001).Biomarker-guided management may substantially shorten the duration of therapy for many patients with MDR-tuberculosis.