Scientists and industry have put significant efforts in creating suitable tools to analyze information flows. However, up to now there are no successful solutions for 1) dynamic modeling of the user-defined interests and further personalization of the results, 2) effective cross-language information retrieval, and 3) processing of multilingual content. As a consequence, much of the potentially relevant and otherwise accessible data from the media stream may elude users’ grasp. We present a multilingual information channeling system, MediaTalk, which offers broad integration between language technologies and advanced data processing algorithms for annotation, analysis and classification of multilingual content. As a result, the system not only provides an all-in-one monitoring service that covers both traditional and social media, but also offers dynamic modeling of user profiles, personalization of obtained data and cross-language information retrieval. Bulgarian and English press clipping services relying on this system implement advanced functionalities such as identification of emerging topics, forecasting and trend prediction, all of which allow the users to monitor their standing reputation, events and relations. The architecture of the system is robust, extensible and adheres to the Big Data paradigm.