Dental microwear texture analysis (DMTA) is a prominent 3D method for investigating diets of extant and extinct species of mammals. Despite its reliability, DMTA faces methodological challenges. Indeed, surfaces with distinct structures can be concealed in the average signal when considering some variables, while they have the potential to discriminate more significantly groups with different diets. To fill this gap, we introduce ‘Trident’, an open-source R package with user interface including a subsampling procedure for performing DMTA. It imports .SUR files, removes aberrant peaks from surfaces, and can subtract polynomial surfaces. Then, it can measure up to 384 variables, corresponding to 24 texture parameters and their statistics (mean, median, standard deviation, skewness…) of distribution over the surface. The computed variables can be classified according to their ability to discriminate discrete categories. At last, ‘Trident’ can display the results in univariate graphs and in multivariate analyses. We here present the use of ‘Trident’ in three case studies, showing how it helps answer questions commonly investigated by paleontologists and archaeologists. First, we separate microwear texture of four controlled-fed groups of pigs based on their diets. Second, we identify discriminant microwear texture parameters in 15 extant species of primates that we relate to biomechanical and ecological factors. Third, we infer the diet of an extinct antelope from the Pleistocene of Greece thanks to a present-day model set up with sympatric European species of ruminants. These case studies show how ‘Trident’ can leverage the results of DMTA.