Tissue plasminogen activator (tPA) is the only FDA approved treatment for ischemic stroke but carries significant risks, including major hemorrhage. Additional options are needed, especially in small vessel thrombi which account for ~25% of ischemic strokes. We have previously shown that tPA-functionalized colloidal microparticles can be assembled into microwheels (µwheels) and manipulated under the control of applied magnetic fields to enable rapid thrombolysis of fibrin gels in microfluidic models of thrombosis. Providing a living microfluidic analog, transparent zebrafish larvae have a highly conserved coagulation cascade that enables studies of hemostasis and thrombosis in the context of intact vasculature, clotting factors, and blood cells. Here we show that tPA-functionalized µwheels can perform rapid and targeted recanalization in vivo. This effect requires both tPA and µwheels, as minimal to no recanalization is achieved with tPA alone, µwheels alone, or tPA-functionalized microparticles in the absence of a magnetic field. We evaluated tPA-µwheels in CRISPR-generated plasminogen (plg) heterozygous and homozygous mutants and confirmed that tPA-µwheels are dose-dependent on plasminogen for lysis. We have found that magnetically powered µwheels as a targeted tPA delivery system are dramatically more efficient at plasmin-mediated thrombolysis than systemic delivery in vivo. Further development of this system in fish and mammalian models could enable a less invasive strategy for alleviating ischemia that is safer than directed thrombectomy or systemic infusion of tPA.