Online test protocols are increasingly popular in psychological and neuroscientific research. Despite its relevance to the social functioning, the influence of acute stress on cognitive and affective state empathy is not clearly understood. Recently, a remote online version (TSST-OL) of the Trier Social Stress Test (TSST) was established for use in research with both children and adults. In general, the TSST-OL offers the opportunity for context-independent application (e.g., at the participants' home or in field contexts). However, in order to exploit this opportunity, it seems crucial to validate the TSST-OL across different settings and contextual variables. We compared stress reactivity in response to the TSST-OL at home and in the laboratory. In a 2 ×2 factorial design, N=120 participants (n=60 women) underwent the TSST-OL and an online adaption of the friendly TSST (fTSST-OL) either at home (n=60) or at the laboratory (n=60). Stress induction was evaluated in terms of physiological (cortisol and salivary alpha-amylase, sAA) and subjective stress and affect measures. Participants also completed an empathy performance task after stress and control exposure. Results confirmed that the TSST-OL successfully induced stress both when conducted at participants' homes and in the laboratory. Still, cortisol levels were higher during laboratory participation compared to application at home, likely due to anticipatory stress. Consequently, the TSST-OL in a home-based application seems to buffer anticipatory stress thus making it an attractive tool to study experimentally induced stress reactivity. Concerning empathy, positive emotions were generally better identified (cognitive empathy) and empathized (affective empathy) than negative emotions. For the latter, this difference was absent after stress, indicated by decreased affective empathy for positive emotions. Overall, this study indicates that the TSST-OL induces stress and validates the tool using a rigorous study design with sufficient participants and relevant stress parameters. Thus, future studies may apply the TSST-OL in different contexts and diverse samples. The findings on empathy under stress align with mixed results in existing research, highlighting the necessity for further investigations into empathy, considering various measurements, stimulus valence, and sex of the participant.