This study examines the effect of biased versus neutral mediation on the content of peace agreements. The author argues that neutral mediators, who are engaged primarily because of their interest to end the war, will have incentives to hasten the reaching of an agreement to the expense of its quality. By contrast, biased mediators, seeking to protect their proteges, will take care to ensure that there are stipulations in an agreement guaranteeing the interest of ‘their’ side or use their particular access and leverage to make their side agree to costly concessions. Biased mediation processes are therefore more likely than neutral mediation processes to lead to elaborated institutional arrangements that are generally considered conducive to democracy and durable peace, such as power sharing, third-party security guarantees, and justice provisions. Empirical analysis, covering the 1989-2004 period and building on data from 124 peace agreements, supports these claims.