This dataset includes MALDI-TOF mass spectra of Klebsiella isolates collected from three Italian clinical centers (University Hospital “Policlinico G. Rodolico”, Catania; “Humanitas” Research Hospital, Rozzano; University Hospital “Foundation Policlinico Gemelli”, Rome), annotated with antimicrobial resistance profiles for meropenem and amikacin. The data support research on machine learning-based AMR prediction from MALDI-TOF spectra with cross-site harmonization. Users are encouraged to process and analyze the data with MaldiAMRKit and combatlearn, which are the tools used for spectral preprocessing and batch-effect correction in the pipeline shared in MALDI-Kleb-AI on GitHub.The dataset, preprocessing pipeline, and machine learning framework used for antimicrobial resistance prediction from MALDI-TOF spectra are described in detail in Combining mass spectrometry and machine learning models for predicting Klebsiella pneumoniae antimicrobial resistance: a multicenter experience from clinical isolates in Italy. Rocchi, E., Nicitra, E., Calvo, M. et al. Combining mass spectrometry and machine learning models for predicting Klebsiella pneumoniae antimicrobial resistance: a multicenter experience from clinical isolates in Italy. BMC Microbiol 26, 180 (2026). https://doi.org/10.1186/s12866-025-04657-2