The past two decades saw several spillovers of animal coronaviruses (aCoVs) to humans. Reporting in Science Immunology, Klompus et al. investigated antibody cross-reactivity between human coronaviruses (hCoVs) and aCoVs by probing the antibody repertoires of 269 individuals who had recovered from COVID-19 against a library of ~13,000 peptides derived from all seven hCoVs and 49 aCoVs. Several broadly reactive monoclonal antibodies showed marked interspecies cross-reactivity. The authors demonstrated that antibody binding data combined with machine learning allows to accurately distinguish between individuals who had been exposed to SARS-CoV-2 and those who had not, even when SARS-CoV-2-specific peptides were excluded from the library. This suggests a potential application of similar antigen libraries for the rapid serological detection of future zoonotic spillovers.