Work-related stress is a major challenge for employee well-being and public health, yet it remains difficult to study as it unfolds in everyday work. Much existing research relies on retrospective self-reports, which are vulnerable to recall bias and often miss situational detail. This study presents the Stress Work Emotion Algorithm, SWEA, Toolbox, a real-time method that combines electrodermal activity, EDA, with event-triggered diary reports to identify and contextualize stress-related arousal in natural work settings. The method was examined across three stages: controlled experimental model development, external benchmarking, and field deployment. By linking physiological signals to participants' brief descriptions of ongoing situations, the toolbox makes it possible to interpret elevated arousal in context rather than treating all arousal as stress by default. The results show that the SWEA Toolbox can generate synchronized physiological and contextual data under controlled and real-world conditions. The experimental stage supported the method's ability to capture a stressrelevant response profile, while the benchmark analysis showed that the event-detection model functioned as a conservative classifier, identifying a meaningful subset of stress-labeled arousal events with very few baseline false alarms. The field studies further showed that the system could detect event-related arousal during ordinary work and produce usable event-linked diary material across settings. The field material also indicated that elevated arousal could reflect both negatively and positively valenced situations, highlighting the importance of contextual annotation for interpretation. The findings suggest that the SWEA Toolbox adds an event-based way to study workplace stress as it unfolds in real-world contexts.