Abstract: A system for automatically enhancing Web pages with annotations expressed in Extensible Markup Language (XML) which describes the pages' multimedia content. Each Web page is parsed or scanned to identify markup tags which contain the URLs of separately stored multimedia data (e.g. image, audio or video files). Each referenced multimedia data entity is then retrieved and analyzed by a type-specific process to extract metadata which describes its content. Additional descriptive metadata may be obtained from the referencing markup tag, accepted from a human editor, or fetched from operating system directories which provide access to the multimedia files. The resulting metadata is expressed in text-based XML format and inserted into a copy of the Web page to form an enhanced Web page whose multimedia content may then be processed by conventional text-based indexing and searching facilities.