Predicting the functional potential of microorganisms in environmental samples from cultivation-independent techniques is a major challenge. A persistent difficulty lies in associating taxonomic profiles obtained from metabarcoding experiment with accurate functional profiles, particularly for poorly-resolved taxonomic groups. In this paper, we present EsMeCaTa a python package predicting shared proteins from taxonomic affiliations. EsMeCaTa relies on the UniProt database to retrieve the public proteomes associated with a taxon and then uses MMseqs2 in order to compute the set of proteins shared in the taxon. Finally, EsMeCaTa extracts the functional annotations of these proteins to provide an accurate estimate of the functional potential associated to taxonomic affiliations.