The core of this project has to do with the development of a clustering mechanism both for texts and for users coming from the internet. We aim towards modeling the navigation paths that common users follow, the automatic evaluation of navigation behaviors with the obvious positive effect being the prediction of future user choices. Within the scope of this mechanism we will be studying classic clustering algorithms and we will be researching the modification and enhancement of those in order to improve their performance under the continuously altering user interests. Additionally, as far as text preprocessing is concerned, we will be researching towards improving it by incorporating query expansion, language thesauri and similar techniques in order to deal with the problems of polysemy and synonymy. User modeling will have a direct impact in our ability of personalizing information for the end user. Furthermore, we will be constructing a personalization algorithm that will be taking into consideration the plethora of parameters which reveal indirectly the user preferences. In the final state of information flow, the results will be returned either to the desktop application or via the Web Interface of our service. The above sub-processes after being evaluated autonomously, will be incorporated in a unified news articles indexing and personalization system.