In this paper, we address the issue of interactive search results manipulation, as provided by typical Web-based information retrieval modules like search engines and directories. Many digital library systems could benefit a lot from the proposed approach, since it is heavily based on metadata, which constitute the building block of such systems. We also propose a way of ranking search results according to their overall importance, which is defined as a weighted combination of the relevancy and popularity of a resource that is being referenced in a search results list. In order to evaluate this model, we have developed an interactive search results manipulation application, which is executed at the client’s workspace through a Web browser without any further interaction with the server that provided the initial search results list. The prototype implementation is based on the XML standard and has been evaluated through an adequate evaluation process from which useful conclusions have been obtained.