Background: The microbiome on dental composites has not been studied in detail before. It has not been conclusively clarified whether restorative materials influence the oral microbiome. Methods: We used Illumina Miseq next-generation sequencing of the 16S V1-V2 region to compare the colonisation patterns of bovine enamel (BE) and the composite materials Grandio Flow (GF) and Grandio Blocs (GB) after 48 h in vivo in 14 volunteers. Applying a new method to maintain the oral microbiome ex vivo for 48 h also, we compared the microbiome on GF alone and with the new antimicrobial substance carolacton (GF+C). Results: All in vitro biofilm communities showed a higher diversity and richness than those grown in vivo but the very different atmospheric conditions must be considered. Contrary to expectations, there were only a few significant differences between BE and the composite materials GB and GF either in vivo or in vitro: Oribacterium, Peptostreptococcaceae [XI][G-1] and Streptococcus mutans were more prevalent and Megasphaera, Prevotella oulorum, Veillonella atypica, V. parvula, Gemella morbillorum, and Fusobacterium periodonticum were less prevalent on BE than on composites. In vivo, such preferences were only significant for Granulicatella adiacens (more prevalent on BE) and Fusobacterium nucleatum subsp. animalis (more prevalent on composites). On DNA sequence level, there were no significant differences between the biofilm communities on GF and GF+C. Conclusion: We found that the oral microbiome showed an increased richness when grown on various composites compared to BE in vitro, but otherwise changed only slightly independent of the in vivo or in vitro condition. Our new ex vivo biofilm model might be useful for pre-clinical testing of preventive strategies.