Diabetes insipidus (DI), be it from central or nephrogenic origin, must be differentiated from secondary forms of hypotonic polyuria such as primary polydipsia. Differentiation is crucial since wrong treatment can have deleterious consequences. Since decades, the gold standard for differentiation has been the water deprivation test, which has limitations leading to an overall unsatisfying diagnostic accuracy. Furthermore, it is cumbersome for patients with a long test duration. Clinical signs and symptoms and MRI characteristics overlap between patients with DI and primary polydipsia. The direct test including vasopressin (AVP) measurement upon osmotic stimulation was meant to overcome these limitations, but failed to enter clinical practice mainly due to technical constraints of the AVP assay. Copeptin is secreted in equimolar amount to AVP but can easily be measured with a sandwich immunoassay. A high correlation between copeptin and AVP has been shown. Accordingly, copeptin mirrors the amount of AVP in the circulation and has led to a ‘revival’ of the direct test in the differential diagnosis of DI. We have shown that a baseline copeptin, without prior thirsting, unequivocally identifies patients with nephrogenic DI. In contrast, for the differentiation between central DI and primary polydipsia, a stimulated copeptin level of 4.9 pmol/L upon hypertonic saline infusion differentiates these two entities with a high diagnostic accuracy and is superior to the water deprivation test. Close sodium monitoring during the test is a prerequisite. Further new test methods are currently evaluated and might provide an even simpler way of differential diagnosis in the future.