Objective: The aim of this study is to evaluate the performance of MALDI-TOF mass spectrometry in identifying bacteria isolated in the oral cavity known to be of probiotic interest. Design: We evaluated Bruker MALDI Biotyper for the identification of 92 clinical oral isolates of probiotic interest (31 Streptococcus salivarius and 61 Lactobacillus spp.) by comparing direct colony method with on-plate formic acid extraction. Isolates were previously identified by use of biochemical methods and molecular biology. Results: Using the manufacturer's suggested genus and species level cutoff scores, the direct colony method identified 42 (45.7%) isolates at the genus level and 35 (38%) at the species level while the on-plate extraction method correctly identified 90 (97.8%) isolates at the genus level and 82 (89.1%) at the species level. The difference between the two methods was statistically significant at the genus and species levels (P ≤ 0.0001). After dividing the isolates into two subgroups, the analysis was repeated. The direct colony method identified correctly all isolates of Streptococcus salivarius at the species level. In contrast, the direct colony method allowed the identification of only 11 (18%) lactobacilli at the genus level and 4 (6.6%) at the species level. The on-plate extraction method was statistically (P ? 0.0001) more efficient since 59 (96.7%) lactobacilli were identified at the genus level and 51 (83.6%) at the species level. Conclusions: MALDI Biotyper can efficiently identify Streptococcus salivarius regardless of the preparative method but on-plate extraction is superior to direct colony method for the identification of lactobacilli.< Réduire