Objective: Periodontal diseases are triggered by dysbiotic microbial biofilms. Therefore, it is essential to develop appropriate biofilm models. Aim of the present study was to culture microcosm biofilms inoculated from different niches in periodontitis patients and compare their microbial composition to those inoculated from subgingival plaque. Methods: Saliva, subgingival plaque, tongue and tonsils were sampled in five periodontitis patients to serve as inocula for culturing biofilms in vitro in an active attachment model. Biofilms were grown for 14 or 28 d and analyzed for their microbial composition by 16S rDNA sequencing. Results: As classified by HOMD, all biofilms were dominated by periodontitis-associated taxa, irrespective which niche had been used for inoculation. There was a low similarity between 14 d biofilms and their respective inocula (Bray-Curtis similarity 0.26), while biofilms cultured for 14 and 28 d shared high similarity (0.69). Principal components analysis showed much stronger clustering per patient than per niche indicating that the choice of patients may be more crucial than choice of the respective niches in these patients. Conclusion: Saliva, tongue scrapings or tonsil swabs may represent sufficient alternative inocula for growing microcosm biofilms resembling periodontitis-associated microbial communities in cases when sampling subgingival plaque is not possible.