<p>Recommender systems are among social media sites’ most critical components in attracting and retaining users. However, they entail challenges, for example, by inducing social comparison, which harms social media users’ well-being. This study developed design knowledge for sensitive social media recommender systems that help to foster users’ well-being. It followed an incremental and iterative design science research approach, evaluating sensitive social media recommender systems through a systematic literature review, two qualitative interview series with experts and users, a scientific focus group, and an online survey based on the Kano customer satisfaction model. The study outcomes include a conceptual framework, meta-requirements, design principles, and design features. This work enhances the understanding of making current social media recommender systems more sensitive towards users, enriches the research on digital responsibility, and is one of the first to demonstrate the feasibility of Kano analysis as an evaluation tool in design science research.</p>