Plant transpiration is the main evaporative flux from terrestrial ecosystems; it controls land surface energy balance, determines catchment hydrological responses and influences regional and global climate. Transpiration regulation by plants is a key (and still not completely understood) process that underlies vegetation drought responses and land evaporative fluxes under global change scenarios. Thermometric methods of sap flow measurement have now been widely used to quantify whole-plant and stand transpiration in forests, shrublands and orchards around the world. A large body of research has applied sap flow methods to analyse seasonal and diurnal patterns of transpiration and to quantify their responses to hydroclimatic variability, but syntheses of sap flow data at regional to global scales are extremely rare. Here we present the SAPFLUXNET initiative, aimed at building the first global database of plant-level sap flow measurements. A preliminary metadata survey launched in December 2015 showed an encouraging response by the sap flow community, with sap flow data sets from field studies representing >160 species and >120 globally distributed sites. The main goal of SAPFLUXNET is to analyse the ecological factors driving plant- and stand-level transpiration. SAPFLUXNET will open promising research avenues at an unprecedented global scope, namely: (i) exploring the spatio-temporal variability of plant transpiration and its relationship with plant and stand attributes, (ii) summarizing physiological regulation of transpiration by means of few water-use traits, usable for land surface models, (iii) improving our understanding of the coordination between gas exchange and plant-level traits (e.g., hydraulics) and (iv) analysing the ecological factors controlling stand transpiration and evapotranspiration partitioning. Finally, SAPFLUXNET can provide a benchmark to test models of physiological controls of transpiration, contributing to improve the accuracy of individual water stress responses, a key element to obtain robust predictions of vegetation responses to climate change.