The UniProt Consortium has updated its protein knowledgebase to provide a comprehensive, high-quality, and freely accessible set of protein sequences annotated with functional information. In 2021, the number of sequences in UniProtKB has risen to approximately 190 million, despite efforts to reduce sequence redundancy. New methods have been adopted to assess proteome completeness and quality. Detailed annotations are extracted from the literature to add to reviewed entries, while unreviewed entries are supplemented with annotations from automated systems such as the Association-Rule-Based Annotator (ARBA). A credit-based publication submission interface allows the community to contribute publications and annotations to UniProt entries. During the COVID-19 pandemic, UniProt responded by curating relevant entries and making them available through a dedicated portal. UniProt resources are available under a CC-BY (4.0) license. UniProt provides a complete compendium of all known protein sequence data linked to functional information. The UniProt Knowledgebase (UniProtKB) combines reviewed entries with unreviewed entries annotated by automated systems. The UniRef databases cluster sequences at various levels of sequence identity, while the UniProt Archive (UniParc) delivers a complete set of known sequences. UniProt integrates and standardizes data from multiple resources to add biological knowledge and metadata to protein records. It is recognized as an ELIXIR Core Data Resource and has received the CoreTrustSeal certification. The data fully supports the FAIR data principles. UniProt continues to evolve to meet new challenges while capturing all available protein sequence data and curating functional data from the scientific literature. The number of entries in UniProtKB has grown by over 65 million in two years, largely due to high-quality metagenomic assembled genomes. The Proteomes portal provides a complete set of proteomes from sequenced genomes, with the BUSCO scoring method used to assess proteome completeness. The Proteomes webpage has been redesigned to allow users to view proteome details in a single table. The 'Complete Proteome Detector' (CPD) algorithm statistically evaluates proteome completeness and quality. Expert curation of experimental data from the scientific literature is fundamental to UniProt. Functional information is added in the form of human-readable summaries and structured vocabularies such as the Gene Ontology. The curators also improve computational accessibility by updating UniProt records with the Rhea knowledgebase of biochemical reactions. Rhea uses the ChEBI ontology to describe reaction participants and their chemical structures. The annotation of pseudoenzymes has been reviewed and updated. Automatic annotation systems, such as the Association-Rule-Based Annotator (ARBA), are used to provide functional annotations for unreviewed entries. These systems use InterPro member databases to classify sequences and predict functional domains. The number of UniRules used for annotation has increased to 6768. ARBA is trained on UniProtKB/Swiss-Prot and generates concise annotation models basedThe UniProt Consortium has updated its protein knowledgebase to provide a comprehensive, high-quality, and freely accessible set of protein sequences annotated with functional information. In 2021, the number of sequences in UniProtKB has risen to approximately 190 million, despite efforts to reduce sequence redundancy. New methods have been adopted to assess proteome completeness and quality. Detailed annotations are extracted from the literature to add to reviewed entries, while unreviewed entries are supplemented with annotations from automated systems such as the Association-Rule-Based Annotator (ARBA). A credit-based publication submission interface allows the community to contribute publications and annotations to UniProt entries. During the COVID-19 pandemic, UniProt responded by curating relevant entries and making them available through a dedicated portal. UniProt resources are available under a CC-BY (4.0) license. UniProt provides a complete compendium of all known protein sequence data linked to functional information. The UniProt Knowledgebase (UniProtKB) combines reviewed entries with unreviewed entries annotated by automated systems. The UniRef databases cluster sequences at various levels of sequence identity, while the UniProt Archive (UniParc) delivers a complete set of known sequences. UniProt integrates and standardizes data from multiple resources to add biological knowledge and metadata to protein records. It is recognized as an ELIXIR Core Data Resource and has received the CoreTrustSeal certification. The data fully supports the FAIR data principles. UniProt continues to evolve to meet new challenges while capturing all available protein sequence data and curating functional data from the scientific literature. The number of entries in UniProtKB has grown by over 65 million in two years, largely due to high-quality metagenomic assembled genomes. The Proteomes portal provides a complete set of proteomes from sequenced genomes, with the BUSCO scoring method used to assess proteome completeness. The Proteomes webpage has been redesigned to allow users to view proteome details in a single table. The 'Complete Proteome Detector' (CPD) algorithm statistically evaluates proteome completeness and quality. Expert curation of experimental data from the scientific literature is fundamental to UniProt. Functional information is added in the form of human-readable summaries and structured vocabularies such as the Gene Ontology. The curators also improve computational accessibility by updating UniProt records with the Rhea knowledgebase of biochemical reactions. Rhea uses the ChEBI ontology to describe reaction participants and their chemical structures. The annotation of pseudoenzymes has been reviewed and updated. Automatic annotation systems, such as the Association-Rule-Based Annotator (ARBA), are used to provide functional annotations for unreviewed entries. These systems use InterPro member databases to classify sequences and predict functional domains. The number of UniRules used for annotation has increased to 6768. ARBA is trained on UniProtKB/Swiss-Prot and generates concise annotation models based