Netwotk_information: Information of every interacting network for which the different specialization metrics were estimated from the repository https://www.web-of-life.es/ (except data of lichens). Contains the number of species of both classes of organisms, the total number of species of each networks, the country and the coordinates. Lichen_networks: Information of every interacting network for mycobiont species and Nostoc phylogroups of 9 forests across a latitudinal gradient in Chile. Each sheet of the Excel document shows the incidence matrix of each of the forest. Abstract: Specialization remains as a controversial and ambiguous term in ecology. Although it has been usually measured using a dichotomic and simplified classification of specialists and generalists, its nature is by far more complex. In the case of biotic interactions, the assignation of these two labels, is usually based on the number of interacting partners (one or few versus many). Here we provide a more precise, quantitative, and objective interpretation of the specialization phenomenon combining three different dimensions and metrics (partner richness, Simpson’s evenness, and d’-index) that offer complementary information to quantify specialization. Hence, partner richness is a metric associated with the specificity, Simpson’s evenness is related with the preference and d’ index with the selectivity of the biotic interactions. Consequently, we propose a 3D specialization space combining these three metrics which allows to identify the degree of biotic specialization fleeing from its simplified historical interpretation. The proposed space was subsequently evaluated in five natural interacting systems (host-parasite, plant-ant, plant-pollinator, plant-disperser, and mycobiont-cyanobacteria) using available data comprising 116 networks with quantitative observations. The results indicate the prevalence of a lax specialization, where most organisms tended to show low values in at least one of the metrics. Predominantly, observations showed high values of specificity and low values of preference and selectivity. This relaxed specialization may provide advantages of being specialized, without sentencing specialization to its limitations when being too tight. The implementation of this framework provides a useful tool that allows to identify specialization in a more objective, integrative, and universal way for future specialization studies.