ChEMBL is a large-scale bioactivity database for drug discovery, containing binding, functional, and ADMET information for a large number of drug-like bioactive compounds. The data are manually abstracted from the primary published literature and curated to ensure quality and utility across various research areas. The database currently contains 5.4 million bioactivity measurements for over 1 million compounds and 5200 protein targets. It is accessible via a web interface, data downloads, and web services at https://www.ebi.ac.uk/chembldb. The database includes data extracted from peer-reviewed publications, with structures of small molecules drawn in machine-readable format. Structures are checked for potential issues and normalized to ensure consistency. Assays include binding, functional, and ADMET assays, with activity endpoints recorded with values and units as given in the paper, and standardized for easier querying. Targets are represented robustly and consistently, with detailed manual annotation to ensure accurate identification and classification. ChEMBL also contains structures and annotations for FDA-approved drugs, including information on trade names, administration routes, dosage, and approval dates. The database includes a data model with entities such as documents, compounds, assays, and targets, with unique identifiers for each. The database is integrated with other resources like PubChem BioAssay and BindingDB, allowing data exchange and complementary information. The database provides various tools for data access, including a user-friendly interface for searching compounds, targets, and assays. Users can retrieve bioactivity data, filter by activity types, and download data for further analysis. The database also includes report cards for approved drugs, providing details on drug type, administration, and safety warnings. ChEMBL is freely available for download in various formats and provides RESTful web services for programmatic access. A subset of the database is available in PSI-MITAB 2.5 format for interoperability with molecular interaction and pathway data. The database is continuously updated and maintained, with a focus on providing high-quality, curated bioactivity data for drug discovery research.ChEMBL is a large-scale bioactivity database for drug discovery, containing binding, functional, and ADMET information for a large number of drug-like bioactive compounds. The data are manually abstracted from the primary published literature and curated to ensure quality and utility across various research areas. The database currently contains 5.4 million bioactivity measurements for over 1 million compounds and 5200 protein targets. It is accessible via a web interface, data downloads, and web services at https://www.ebi.ac.uk/chembldb. The database includes data extracted from peer-reviewed publications, with structures of small molecules drawn in machine-readable format. Structures are checked for potential issues and normalized to ensure consistency. Assays include binding, functional, and ADMET assays, with activity endpoints recorded with values and units as given in the paper, and standardized for easier querying. Targets are represented robustly and consistently, with detailed manual annotation to ensure accurate identification and classification. ChEMBL also contains structures and annotations for FDA-approved drugs, including information on trade names, administration routes, dosage, and approval dates. The database includes a data model with entities such as documents, compounds, assays, and targets, with unique identifiers for each. The database is integrated with other resources like PubChem BioAssay and BindingDB, allowing data exchange and complementary information. The database provides various tools for data access, including a user-friendly interface for searching compounds, targets, and assays. Users can retrieve bioactivity data, filter by activity types, and download data for further analysis. The database also includes report cards for approved drugs, providing details on drug type, administration, and safety warnings. ChEMBL is freely available for download in various formats and provides RESTful web services for programmatic access. A subset of the database is available in PSI-MITAB 2.5 format for interoperability with molecular interaction and pathway data. The database is continuously updated and maintained, with a focus on providing high-quality, curated bioactivity data for drug discovery research.