Coffee byproducts, such as coffee leaves, are components of the coffee plant that are still underexplored. Considering their potential application in the food chain, defining their phytochemical profile and its susceptibility to processing is essential. A comprehensive HR-MS profiling performed using ultrahigh-performance liquid chromatography coupled with traveling wave ion mobility spectrometry/quadrupole time-of-flight mass spectrometry was applied to dive into the phytochemicals occurring in Coffea arabica L. leaves (cv. red Caturra), following three postharvest approaches used to obtain an oxidized, a slightly oxidized, and a nonoxidized, herbal infusion-like botanical product. Coffee signature compounds, such as caffeine and 5-caffeoylquinic acid, were also quantified, evidencing that each processing step had a peculiar impact on the extracts. Slightly oxidized and oxidized samples showed comparable levels of caffeine and 5-caffeoylquinic acid, while the content of both compounds was significantly different in nonoxidized leaves. The untargeted HR-MS approach followed by chemometrics allowed a clear clustering of the three different treatments, and 472 identifications were putatively assigned. The large majority (95.4%) of the identified metabolites fall within 8 chemical superclasses, the most represented being phenylpropanoids (38.0%), sterol lipids (18.2%), and glycerophospholipids (17.7%). Among all of the identified metabolites, 45% are significantly overaccumulated in nonoxidized samples, 37% in the oxidized samples, and, finally, 18% in the slightly oxidized samples. Besides providing exhaustive profiling, such an approach confirms the suitability of untargeted HR-MS for discrimination of commercial batches of coffee leaves and highlights how its exploitation could support future investigations.