The wide diffusion of community tagging sites and related folksonomies has made the knowledge discovery and retrieval still much more urgent topic. If tagging systems allow users to add freely keywords to web resources, clicking on a tag has the side effect of a tag-based query, since enables the users to explore related content. The collective knowledge expressed though user-annotated data has a big potential, but needs to be filtered in a digest form so that the search result better reflects the users' preferences and actual aims. Starting from these considerations, our work presents an agent-based approach for a scalable semi-automatic generation of annotation tags, personalized on each user's preferences and tastes. Primary is the role of agents which assist users in the tagging activities as well as the retrieval of resources related to their interest. A user-friendly interface proposes an integrated one-shot view for interacting with a tagging system.
