Within the emergent Semantic Web framework, the use of traditional web search engines based on keywords provided by the users is not adequate anymore. Instead, new methods based on the semantics of user keywords must be defined to search in the vast Web space without incurring in an undesirable loss of information.

In this paper we propose a system that takes as input a list of plain keywords provided by the user and outputs equivalent semantic queries expressed in a knowledge representation language, that could be used to retrieve relevant data. For the translation task, specialized agents manage a third-party thesaurus and a pool of pre-existing ontologies to obtain the different meanings of the user keywords and discover semantic relationships between them in run-time.