The emergence of social commerce has brought substantial changes to both businesses and consumers. Amid this backdrop, understanding consumer behavior in social commerce contexts is critical to sellers that endeavor to more effectively influence consumers and capitalize on the power of social ties. With the technical features of social commerce website and the stimulus-organism-response paradigm as bases, this study develops a model to investigate the effects of technical features (interactivity, recommendations, and feedback) on relationship quality (swift guanxi and trust) and subsequent repurchase intention. Collecting 506 valid respondents of agricultural product consumers in social commerce, we utilized SmartPLS to conduct statistical analysis for the model. The empirical results indicate that interactivity, recommendations, and feedback exert positive effects on swift guanxi and trust to different degrees. In turn, swift guanxi and trust enable and mediate the prediction of consumer repurchase intention in social commerce context. The theoretical and pragmatic implications for firms in social commerce market are also provided.