A better understanding of the extent to which coastal species passively disperse and establish connections across seascapes is instrumental for marine conservation and management, as well as biogeography and climate change research. This understanding is particularly relevant for projecting the dynamics of species' geographical ranges. However, a comprehensive perspective on the elemental structures and patterns of species connectivity along coastlines is currently lacking. To address this gap, we present a mechanistic biophysical approach designed to reveal communities in connectivity networks for a broad spectrum of taxonomic groups, including fishes, corals, molluscs, crustaceans, macroalgae, seagrasses, and mangroves on a global scale. This approach is grounded by 21 years of ocean current data and tuned by empirical information on species dispersal capacity, with kernels spanning from local to long-distance dispersal events. We delineate well-connected communities as distinct oceanographic units, explaining observed community composition derived from species range maps (approximately 4,000 species globally). We identify the location of connectivity barriers and the directional putative connectivity pathways at the global scale, revealing a global network between oceanographic units for each taxonomic group under consideration. By expanding connectivity analysis to the biogeographic scale, our findings establish baselines to enhance our understanding of the mechanisms controlling species dispersal and ongoing range shifts due to climate change. Additionally, they provide standards for the development of effective conservation and management spatial strategies for marine coastal species.