AIM: To evaluate in vitro the antibacterial efficacy and cytocompatibility of different implant-decontamination methods, using both 2D and 3D peri-implant mucosa models. METHODS: Four decontamination methods [chlorhexidine (CHX), electrolytic treatment (GS), curcumin (CUR), xanthohumol (XN)] were compared in four independent experiments, three with a 2D peri-implant mucosa model on titanium surfaces and another on a 3D peri-implant mucosa model. These decontamination procedures were tested for their antibacterial effect using a multispecies biofilm model with Streptococcus oralis, Actinomyces naeslundii, Veillonella dispar, and Porphyromonas gingivalis for 24 h. Direct cytocompatibility evaluating the impact of the treatments on tissue cells as well as indirect cytocompatibility (colonization of treated implant surfaces by tissue cells) were also tested. Both outcomes were assessed by confocal laser scanning microscopy supported by neural networks. RESULTS: CHX demonstrated a strong alteration of cytocompatibility and antibacterial effect, but did not remove biofilm biomass. XN and CUR demonstrated antibacterial effect and biofilm removal action, as well as cytocompatibility. GS showed antibacterial capacity with a combination of areas completely clean of biofilm with others in which a non-vital biofilm remained. In the 3D peri-implant mucosa model, XN and CUR showed maintenance of the mucosa integrity after treatment, whereas CHX and GS displayed disruption in the mucosal layers. CONCLUSIONS: Phytotherapeutics (CUR and XN) were the most cytocompatible substances and showed the largest antimicrobial effect. GS displayed antibiofilm activity with a localized "bubble-shaped effect" and impaired tissue cell morphology and integrity, compromising cytocompatibility, and CHX showed antimicrobial capacity, without reducing biofilm biomass and with altered cytocompatibility.