PeptDeep provides deep learning models for mass spectrometry-based proteomics. It includes built-in models for predicting retention time, collision cross section, and tandem mass spectra for peptides, enabling users to generate predicted spectral libraries from protein sequences.