Cyanobacteria and algae living inside carbonate rocks (endoliths) have long been considered major contributors to bioerosion. Some bore into carbonates actively (euendoliths); others simply inhabit pre-existing pore spaces (cryptoendoliths). While naturalistic descriptions based on morphological identification have traditionally driven the field, modern microbial ecology has shown that this approach is insufficient to assess microbial diversity or make functional inferences. I examined endolithic microbiomes using 16S rRNA genes and lipid-soluble photosynthetic pigments as biomarkers, with the goal of reassessing endolith diversity by contrasting traditional and molecular approaches. This led to the unexpected finding that in all 41 littoral carbonate microbiomes investigated around Isla de Mona (Puerto Rico) and Menorca (Spain) populations of anoxygenic phototrophic bacteria (APBs) in the phyla Chloroflexi and Proteobacteria, were abundant, even sometimes dominant over cyanobacteria. This was not only novel, but it suggested that APBs may have been previously misidentified as morphologically similar cyanobacteria, and opened questions about their potential role as euendoliths. To test the euendolithic role of photosynthetic microbes, I set a time-course experiment exposing virgin non-porous carbonate substrate in situ, under the hypothesis that only euendoliths would be able to initially colonize it. This revealed that endolithic microbiomes, similar in biomass to those of mature natural communities, developed within nine months of exposure. And yet, APB populations were still marginal after this period, suggesting that they are secondary colonizers and not euendolithic. However, elucidating colonization dynamics to a sufficiently accurate level of molecular identification among cyanobacteria required the development of a curated cyanobacterial 16S rRNA gene reference database and web tool, Cydrasil. I could then detect that the pioneer euendoliths were in a novel cyanobacterial clade (named UBC), immediately followed by cyanobacteria assignable to known euendoliths. However, as bioerosion proceeded, a diverse set of likely cryptoendolithic cyanobacteria colonized the resulting pore spaces, displacing euendoliths. Endolithic colonization dynamics are thus swift but complex, and involve functionally diverse agents, only some of which are euendoliths. My work contributes a phylogenetically sound, functionally more defined understanding of the carbonate endolithic microbiome, and more specifically, Cydrasil provides a user-friendly framework to routinely move beyond morphology-based cyanobacterial systematics.