eggNOG-mapper is a tool for fast functional annotation of large sets of sequences based on orthology assignments using precomputed clusters and phylogenies from the eggNOG database. It was validated against two homology-based approaches, BLAST and InterProScan, and showed improved accuracy and efficiency. Orthology filters reduced false positive assignments by 11% and increased the ratio of experimentally validated terms by 15%. Compared to InterProScan, eggNOG-mapper predicted 41 more terms per protein and increased the rate of experimentally validated terms by 35%. It performed within the top-5 methods in the three GO categories using the CAFA2 NK-partial benchmark. eggNOG-mapper runs 15 times faster than BLAST and at least 2.5 times faster than InterProScan. It is available as a standalone package and online service. The tool uses HMMER or DIAMOND for sequence mapping, orthology assignment based on precomputed phylogenetic trees, and functional annotation based on curated GO terms, KEGG pathways, and COG functional categories. It allows for taxonomic filtering and can restrict annotations to one-to-one orthologs. It was tested on five model organisms and showed improved performance compared to BLAST and InterProScan. It also performed well in metagenomics data annotation, yielding better performance than InterProScan. eggNOG-mapper is faster than BLAST and InterProScan, and is suitable for large-scale annotation projects such as metagenomics. It is synchronized with the eggNOG database, ensuring up-to-date annotations. The tool is available for use and can be integrated into third-party bioinformatics pipelines.eggNOG-mapper is a tool for fast functional annotation of large sets of sequences based on orthology assignments using precomputed clusters and phylogenies from the eggNOG database. It was validated against two homology-based approaches, BLAST and InterProScan, and showed improved accuracy and efficiency. Orthology filters reduced false positive assignments by 11% and increased the ratio of experimentally validated terms by 15%. Compared to InterProScan, eggNOG-mapper predicted 41 more terms per protein and increased the rate of experimentally validated terms by 35%. It performed within the top-5 methods in the three GO categories using the CAFA2 NK-partial benchmark. eggNOG-mapper runs 15 times faster than BLAST and at least 2.5 times faster than InterProScan. It is available as a standalone package and online service. The tool uses HMMER or DIAMOND for sequence mapping, orthology assignment based on precomputed phylogenetic trees, and functional annotation based on curated GO terms, KEGG pathways, and COG functional categories. It allows for taxonomic filtering and can restrict annotations to one-to-one orthologs. It was tested on five model organisms and showed improved performance compared to BLAST and InterProScan. It also performed well in metagenomics data annotation, yielding better performance than InterProScan. eggNOG-mapper is faster than BLAST and InterProScan, and is suitable for large-scale annotation projects such as metagenomics. It is synchronized with the eggNOG database, ensuring up-to-date annotations. The tool is available for use and can be integrated into third-party bioinformatics pipelines.