Chronic pelvic pain disorders (CPPDs) affect millions of women and are associated with persistent pain, reduced physical functioning, and high healthcare utilization. Although physical exercise is recommended for symptom management, there are currently no guidelines for optimal pain management in CPPDs. WorkoutCPP is a 9-week, N-of-1 personalized study built on the "ehive" mobile health platform, that uses reinforcement learning (RL) to adapt daily exercise recommendations to address this gap. Study participants track daily symptoms, quality of life, and relevant behaviors using "ehive", and wear an activity tracker for 9 weeks. These variables then inform the dynamic RL model updates. After a warm-up week of generic recommendations, the RL agent updates action probabilities using Bayesian sampling informed by prior exercise type, intensity, duration, and next-day pain reduction. The model continuously balances exploration and exploitation to identify prescriptions associated with improved outcomes while operating within predefined safety and preference constraints. Recommendations are sent through automated notifications to a participant’s smartphone, enabling real-time, AI-driven clinical decision support within a clinician-supervised framework. This study is the first of its kind to investigate AI-based personalized exercise for disease management in a gynecological disease setting. Findings will support global initiatives on nonpharmacologic pain management modalities.