Bootstrapping an Online News Knowledge Base

News retrieval systems facilitate the process of quickly learning about events or stories reported in various online news providers. The traditional approach involves clustering articles that report about the same event using bag-of-words or concept based similarity measures, and offering personalized recommendations using various user modeling approaches. Knowledge bases have been extensively used in the recent years for powering search engines on entity based searches. The success of this approach, demonstrated by a now de-facto way of searching and browsing offered by commercial search engines and mobile applications, has created the need to incorporate semantic capabilities to news retrieval systems. In this paper we present a proposal for creating a knowledge base of entities, events and facts reported in Albanian online news providers. We aim to provide a news stream processing pipeline based in generally available open source toolkits and state-of-the-art research works about event and fact oriented knowledge bases.