This paper presents a new Wiki called SAVVY Wiki that realizes context-oriented, collective and collaborative knowledge management environments that are able to reflect users' intentions and recognitions. Users can collaboratively organize fragmentary knowledge with the help of the SAVVY Wiki. Fragmentary knowledge, in this case, implies existing Wiki content, multimedia content on the web, and so on. Users select and allocate fragmentary knowledge in different contexts onto the SAVVY Wiki. Owing to this operation, it is ensured that related pages belong to the same contexts. That is, users can find correlations among the pages in a Wiki. The SAVVY Wiki provides new collective knowledge created from fragmentary knowledge, depending on contexts, in accordance with the users' collaborative operations. Various collaborative working environments have been developed for the sharing of collective knowledge. Most current Wikis have a collaborative editing mode to every page, as a platform to enable a collaborative working environment. In order to understand an arbitrary concept thoroughly, it is necessary to find correlations among the various threads of content, depending on the users' purpose, task or interest. In a Wiki system, it is important to realize a collaborative editing environment with correlation among pages depending on the contexts. In this paper, we present a method to realize the SAVVY Wiki, and describe its developing prototype system.