Nonviolent protest movements have been prevalent in the last decades. While such movements aim for peaceful change, they are frequently followed by civil war. Previous research has shown that outcomes of nonviolent protests can be influenced by mediation, but because most previous research on conflict mediation has predominately examined armed conflicts, little is known about when and how mediation occurs. We argue that mediation in nonviolent uprisings is more likely when social conflicts generate negative externalities for the outside world, incentivizing third parties to act and conflict parties to accept their terms. After assessing the scope of the empirical field and identifying anchoring points for future research, we examine data on nonviolent campaigns between 1970 and 2014, investigating patterns in mediation incidence across time and space by situational characteristics, and by the origins of the mediator. We find that protest movements with a higher risk of violent escalation, marked by radicalism or state repression, are more likely to be mediated, and that mediation of nonviolent disputes has shifted from domestic to international mediators. We conclude by discussing theoretical implications for the field as well as suggesting some important policy and practice implications for the mediation of nonviolent conflicts.