Abstract Antagonists of platelet integrin alphaIIbbeta3 are potent anti-thrombotics due to critical roles of integrins in thrombosis. However, integrins are also important in hemostasis, and thus integrin antagonists have potentially life-threatening bleeding side effect. It would be ideal if we can develop integrin antagonists without bleeding side effect. Integrins transmit signals bidirectionally. Intracellular signals activate integrin alphaIIbbeta3, leading to talin-dependent integrin ligation, which is required for platelet adhesion and initial hemostatic thrombus formation. Integrin ligation in turn mediates Galpha13/Src-dependent outside-in signaling that stabilizes and amplify thrombi, which is crucial for occlusive arterial thrombosis. Here we show that talin and Galpha13 bind to mutually exclusive but distinct sites in integrin beta3, and their bindings are dynamically regulated during integrin signaling. The first talin binding wave mediates inside-out signaling and also “ligand-induced integrin activation”, but is not required for early phase outside-in signaling. Integrin ligation induces talin dissociation and Galpha13 binding, which selectively mediates early phase outside-in signaling. The second talin binding wave is associated with late phase outside-in signaling and clot retraction. Based on these findings, we have designed a selective inhibitor of Galpha13-integrin interaction, which specifically abolished outside-in signaling without affecting inside-out signaling and integrin ligation. Strikingly, this inhibitor potently inhibits occlusive thrombosis in vivo, but has no effect on tail bleeding time, in contrast to the current integrin antagonists. Thus, we have discovered a mechanism for switching the direction of integrin signaling and a new anti-thrombotic that does not cause bleeding. Disclosures: No elevant conflicts of interest to declare.