It has long been established, that views to natural scenes can a have a dampening effect on physiological stress responses. However, as people in Europe, Canada and North America today spent 50-85% of their time indoors, attention might also be paid to how the artificial man-made indoor environment influences these mechanisms. The question that this study attempts to start addressing is therefore whether certain design, characteristics of indoor spaces can make a difference to the physiological stress response as well. Using a virtual version of the Trier Social Stress Test, in which the space is computer generated and properties of the space therefore can be systematically varied, we measured saliva cortisol and heart rate variability in participants in a closed room versus a room with openings. As shown by a significant linear contrast interaction between groups and TSST conditions, participants in the closed room responded with more pronounced cortisol reactivity to stress induction, and continued to show higher levels throughout recovery, compared to participants in the open room. No differences were found regarding any part of the autonomic nervous system.