Traditional peer-to-peer Supplemental Instruction (SI) was introduced into higher education over a quarter of a century ago and promptly became an integral part of the developmental mathematics curricula in many senior and community colleges. Later, some colleges introduced Video-based Supplemental Instruction (VSI) and, in recent years, Web-delivered Supplemental Instruction (WdSI), enhancing the delivery of SI. A major shortcoming of these new approaches is the fixed content of a prerecorded SI and its restricted availability. Furthermore, frequent changes in curricula and new or revised textbooks can quickly make SI incomplete. This research suggests a novel approach to a Web-delivered SI (WdSI), where the content and elucidation level of the SI are based on the user-supplied topic description and are precisely selected from an enormous, continuously expanding and constantly available Internet collection. A meta search engine with dynamically adjustable search algorithms was designed and used to conduct experiments; the results demonstrated a high quality of on-demand delivery of relevant SI for any discipline or subject matter.