All thrombolytic agents in current clinical usage are plasminogen activators. Although effective, plasminogen activators uniformly increase the risk of bleeding complications, especially intracranial hemorrhage, and no laboratory test is applicable to avoid such bleeding. We report results of a randomized, blinded, dose-ranging comparison of tissue-type plasminogen activator (TPA) with a direct-acting thrombolytic agent, plasmin, in an animal model of fibrinolytic hemorrhage. This study focuses on the role of plasma coagulation factors in hemostatic competence. Plasmin at 4-fold, 6-fold, and 8-fold the thrombolytic dose (1 mg/kg) induced a dose-dependent effect on coagulation, depleting antiplasmin activity completely, then degrading fibrinogen and factor VIII. However, even with complete consumption of antiplasmin and decreases in fibrinogen and factor VIII to 20% of initial activity, excessive bleeding did not occur. Bleeding occurred only at 8-fold the thrombolytic dose, on complete depletion of fibrinogen and factor VIII, manifest as prolonged primary bleeding, but with minimal effect on stable hemostatic sites. Although TPA had minimal effect on coagulation, hemostasis was disrupted in a dose-dependent manner, even at 25% of the thrombolytic dose (1 mg/kg), manifest as rebleeding from hemostatically stable ear puncture sites. Plasmin degrades plasma fibrinogen and factor VIII in a dose-dependent manner, but it does not disrupt hemostasis until clotting factors are completely depleted, at an 8-fold higher dose than is needed for thrombolysis. Plasmin has a 6-fold margin of safety, in contrast with TPA, which causes hemorrhage at thrombolytic dosages.