Identifying the source of seafood is critical for combatting seafood fraud, but current tools are predominantly developed and applied on a species-specific basis. This study investigates how multiple marine taxa could be geolocated at global scales by exploiting stable oxygen isotope compositions in carbonate biominerals (δ 18 O biomin ), where we expect to see universally expressed and predictable spatial variation in δ 18 O biomin values across taxa. We constructed global ocean isoscapes of predicted δ 18 O biomin values specific to fish (otoliths), cephalopod (statoliths) and shellfish (shells), and a fourth combined "universal" isoscape, and evaluated their capacity to derive δ 18 O biomin values among known-origin samples. High correspondence between isoscape-predicted δ 18 O biomin values and a compiled database of measured, georeferenced values (3954 datapoints encompassing 68 species) indicated that this δ 18 O biomin approach works effectively, particularly in regions with highly resolved projections of seawater δ 18 O composition. When compared to taxon-specific isoscapes, the universal isoscape demonstrated similar accuracy, indicating exciting potential for universal provenance applications. We tested the universal framework via a case study, using machinelearning models to identify sample origins amongst regions of divergent (Tropical Asia vs Temperate Australasia) and similar (Temperate Asia vs Temperate Australasia) climates and latitudes. Classification accuracy averaged 75.3% between divergent regions, and 66% between similar regions. When endothermic tuna species were excluded from the analysis, the accuracy between divergent regions increased up to 90% between divergent regions. This study presents the first empirical step towards developing universal chemical markers, which have the potential to support a more inclusive and global approach of validating provenance of seafood.