Abstract
Ontologies have shown to be a key mechanism to deal with a set of different systems. This paper proposes a model of software agents in charged of information retrieval tasks from a federation of digital libraries under Open Archives Initiative Protocol for Metadata Harvesting. We have termed these assistants ontological agents because they operate with ontologies. Ontologies are used to support two main tasks: to eliminate ambiguity from a query and to enable the access to the repositories. We present their model which is based on MAS-CommonKADS methodology.