The paper presents the design plan for a mobile solution aimed at stress reduction. The solution will be developed by a team of medics, psychotherapists, HCI experts and knowledge engineers and will provide continuous data sensing and feedback about personal stress levels. At the same time contextual and activity information will be captured. Stress management is particularly important for high-risk populations such as former alcoholics to reduce the risk of relapse; they will therefore test and validate the solution. By combining and correlating psycho-physiological data with data on activities (e.g. walking or social interactions) and environment/location (e.g. ambient light) it is expected that sources of stress can be recognised which in turn will allow individuals to either avoid stress-inducing factors or develop appropriate coping strategies. To make sense of the data captured, it is proposed to use intelligent algorithms to recognise patterns in the data streams and semantic technologies to interpret the text messages of users. People with other stress-related health problems such as burn-out, smoking, depression or sleeping problems will also benefit from our research.