CRISP (Computer Retrieval of Information on Scientific Projects) is a large database maintained and operated by the National Institutes of Health (NIH). It contains comprehensive scientific and selected administrative data on research carried out by the U.S. Public Health Service (PHS) or supported by PHS grants and contracts. Developed originally to meet the needs of NIH, it is an excellent, largely untapped resource for health information professionals at large, revealing new trends, methods, and techniques, often before they appear in the published literature. CRISP uses its own controlled vocabulary, developed to permit indexing of new and active research areas. Queries can combine subject headings with a great variety of administrative data elements (e.g., research category or principal investigator's name). Output is available in a variety of formats and media. While information professionals cannot directly access the CRISP system, abridged CRISP records are merged into the FEDRIP (Federal Research in Progress) database, and FEDRIP is publicly accessible through DIALOG. CRISP records in toxicology are also furnished to the National Library of Medicine's TOXLINE database. This paper discusses the indexing, information retrieval, publication products, and search services of the CRISP system, and how users of medical information can benefit from it.