The goal of NoTube is to develop a service architecture, based on semantic technologies, for personalised creation, distribution and consumption of TV content in all its new forms. Nowadays, consumers can receive hundreds of digital TV channels. TV content can be rendered on different devices, including handhelds and computers. Through IPTV the TV and the Web worlds are getting intertwined. Companies are starting to supply new services, such as electronic program guides, which recommend broadcasts based on your viewing habits. However, such interactive services stay TV-centric and use Internet and digital technology only to enhance the entertainment in front of the tube. NoTube intends to put the TV viewer back in the driver's seat by developing services for (1) personalized content selection, (2) packaging of content with additional information (e.g. relevant Web info), and (3) social interaction in consumer communities. We build on information integration and representation techniques developed within the Semantic Web area, taking into account the context of time, space, task and mood. To realize this vision NoTube will deliver a service-oriented architecture with four corner stones: (1) novel deployment of existing semantic technologies in the combined TV-Internet environment; (2) extensions of existing user and context modeling techniques to meet the demands of the distributed world; (3) development of novel reasoning services for personalized content recommendation; and (4) integration with social interaction tools. The work in the project is steered by three visionary use cases, namely: (i) personalized semantic news, (ii) personalized TV guide with adaptive advertising, and (iii) Internet TV in the Social Web (for Olympic Games 2012). The use cases address different dimensions of personalized TV-content interaction, including individuals and communities of viewers as well as multi-lingual and multi-modal interaction.