Although previous research has established that interpersonal trust and system trust are critical in shaping individual behaviour in virtual settings, the two perspectives have not been examined by IS researchers in virtual communities (VCs) simultaneously. Drawing from prior literature on trust and VCs, a research model for understanding the importance of trust in members and trust in system and their antecedents in VCs is presented. Six trust-building factors were identified using three trust-building bases (calculative base, relationship base and technology base). Data were collected from 324 members of a technical VC to test the model. The study shows that trust in members and trust in system have significant influences on knowledge sharing intention. The study also indicates that knowledge growth, perceived responsiveness and shared vision affect trust in members positively, while knowledge quality influences trust in system significantly. The study discusses the theoretical and managerial implications of this study and proposes several future research directions. (Contains 2 figures and 5 tables.)