The VISOR project aims to provide virtual coaching services for the elderly aiming to preserve the well-being of senior citizens towards a healthy lifestyle. It will provide personalized motivating recommendations through a mobile application that will run on a low-cost smartphone. Several low-cost sensors (such as a smart-watch or bracelet, cameras, microphones and environmental sensors) will be used to acquire information about the user’s physiological and emotional status and her/his home environment. Using all available sensory information, behavioural patterns, health issues, emotions and simple events will be recognized using various machine learning algorithms and coaching will be realized by an expert system. An Internet-of-Things architecture will be followed. Thus, all sensors and processing algorithms will be implemented as exposed web services running on the cloud. Emphasis will be given to ensure data protection and privacy preservation, thus from the sensitive services, i.e., the visual and audio ones, only low- or mid-level features will be transmitted outside the home environment. Coaching will include recommendations for physical, cognitive and social well-being, while it will be aware for certain emergency issues, such as falls, cardiac problems etc. Evaluation will focus on proof-of-concept and will consider both real-life users and a real-like environment.