Case Based Reasoning (CBR) is a Knowledge Management approach that consists in the development of decision support systems where problem are solved by analogy with similar problem solved in the past. In this way, the system supports users in finding solutions without starting from scratch. CBR has become a very important research topic in Artificial Intelligence, with the definition of methodologies and architectural patterns for supporting developers in the design and implementation of case–based systems. The paper presents one of this frameworks, namely CReP, an on–going research project of the Artificial Intelligence Laboratory (L.Int.Ar.) of University of Milan–Bicocca, focusing on the integration between CBR paradigm and metadata approach to obtain domain–independent case structure and retrieval algorithm definition.