eggNOG-mapper v2 is an updated version of the tool for functional annotation, orthology assignment, and domain prediction at the metagenomic scale. It is optimized for large (meta)genomic datasets and includes several improvements over version 1. The genomes and functional databases have been updated to those from eggNOG v5, and the tool now allows for de novo gene prediction from raw contigs, built-in pairwise orthology prediction, fast protein domain discovery, and automated GFF decoration. eggNOG-mapper v2 is available as a standalone tool or online service at http://eggnogmapper.embl.de. The underlying genome database has been updated to include 4.4 million OGs and more than twice the number of organisms compared to the previous version. This improvement increases annotation coverage and phylogenetic resolution, particularly noticeable when analyzing large metagenomic datasets. The annotation rate has improved by 16% on average, and the annotation rates have increased by 608% compared to version 1. eggNOG-mapper v2 is faster than Prokka, especially on large metagenomic datasets. The new version also allows for the prediction of ORFs directly from assembled contigs using Prodigal. Additionally, the initial sequence-mapping step now supports Diamond, MMseqs2, and HMMER3 modes, each recommended for different use cases. The new Diamond mode '—iterate' iteratively applies different levels of sensitivity. eggNOG-mapper v2 provides new annotation sources and improved reports, including predicted protein names, KEGG pathways, Gene Ontology labels, EC numbers, BiGG reactions, CAZy terms, COG functional categories, eggNOG OGs, and free text descriptions. It also provides PFAM and SMART protein domain predictions, with options for de novo domain annotation. Compared to other functional annotation tools, eggNOG-mapper v2 offers a more efficient, versatile, and scalable automated functional annotation workflow. It is optimized for large-scale annotation jobs and produces GFF files decorated with functional annotations, tab-delimited files with predicted functional terms per query, orthology assignment tables, and taxonomic placements. It is available at GitHub and as an online service.eggNOG-mapper v2 is an updated version of the tool for functional annotation, orthology assignment, and domain prediction at the metagenomic scale. It is optimized for large (meta)genomic datasets and includes several improvements over version 1. The genomes and functional databases have been updated to those from eggNOG v5, and the tool now allows for de novo gene prediction from raw contigs, built-in pairwise orthology prediction, fast protein domain discovery, and automated GFF decoration. eggNOG-mapper v2 is available as a standalone tool or online service at http://eggnogmapper.embl.de. The underlying genome database has been updated to include 4.4 million OGs and more than twice the number of organisms compared to the previous version. This improvement increases annotation coverage and phylogenetic resolution, particularly noticeable when analyzing large metagenomic datasets. The annotation rate has improved by 16% on average, and the annotation rates have increased by 608% compared to version 1. eggNOG-mapper v2 is faster than Prokka, especially on large metagenomic datasets. The new version also allows for the prediction of ORFs directly from assembled contigs using Prodigal. Additionally, the initial sequence-mapping step now supports Diamond, MMseqs2, and HMMER3 modes, each recommended for different use cases. The new Diamond mode '—iterate' iteratively applies different levels of sensitivity. eggNOG-mapper v2 provides new annotation sources and improved reports, including predicted protein names, KEGG pathways, Gene Ontology labels, EC numbers, BiGG reactions, CAZy terms, COG functional categories, eggNOG OGs, and free text descriptions. It also provides PFAM and SMART protein domain predictions, with options for de novo domain annotation. Compared to other functional annotation tools, eggNOG-mapper v2 offers a more efficient, versatile, and scalable automated functional annotation workflow. It is optimized for large-scale annotation jobs and produces GFF files decorated with functional annotations, tab-delimited files with predicted functional terms per query, orthology assignment tables, and taxonomic placements. It is available at GitHub and as an online service.