The inability to discover the conversations that matter, to understand how others are feeling, to recognise a friend in need, to share the joy of someone happy, or to see how certain things should be done, are common limitations of social networks. All these require people to be able to see and hear each other. Vconect targets this space by aiming to develop capabilities that will allow ad-hoc groups of people to enjoy real-time high-quality audio-video communication. Such capabilities will be able to deliver the best possible communication experience to all the participants involved, no matter their number, the types of devices or the characteristics of the bandwidth they are connecting through.The Vconect platform will be able to support complex and dynamic structures of group communication by intelligently mediating the conversation and framing it in narrative structures through methods akin to those used in cinematography. Active cameras, shot types, screen compositions will be decided dynamically, to ensure meaningful, engaging and immersive communication experiences. In turn, these decisions will be optimised in real-time for the network and devices that are being used. The platform thus becomes service-aware, displaying intelligent behaviour both in communication mediation as well as in adapting the way the audio and video is processed within the network to best support the communication needs.This new development builds upon two of the most significant achievements of the current Internet: video conferencing and social networks. It will make social networks as flexible and as engaging as chatting face to face to a group of friends: it will allow us to see what's really happening, to know who's hurting and who's laughing. It will allow us to see the real drama, let us be part of the most rowdy crowds or talk quietly to a lonely friend. It will strengthen communities and change the way we live.