Thrombosis is the underlying pathology of stroke, which is a leading cause of death and disability in developed countries.1 Thrombus formation is controlled by the coagulation cascade, which is made up of two distinct pathways, the intrinsic and extrinsic pathways, which converge on a central path. Current anticoagulant treatments target thrombin, an enzyme in the central path. While thrombin inhibitors have proven to be effective for certain indications, bleeding side effects are a common problem, thus efficacious drugs without these side effects are desirable. FXIIa is a serine protease that directly activates only the intrinsic pathway of coagulation. Murine model studies have shown that deficiency or inhibition of FXIIa leads to normal haemostasis and protection from stroke.2 Cyclic peptides are attractive molecules due to their rigid structure and the possibility of high binding affinity and target selectivity. Random nonstandard Peptide Integrated Discovery (RaPID) is platform technology which can be used to quickly identify cyclic peptides with high binding affinity for a particular target.3 We have identified a library of cyclic peptides with high binding affinity for FXIIa using RaPID. These were synthesised via solid-phase peptide synthesis, cyclising through a cysteine residue onto an N-terminal chloroacetylated amino acid. The synthetic cyclic peptides were assessed in an in vitro FXIIa inhibition assay and shown to have nanomolar inhibitory activity. Screening against other blood coagulation proteases also showed that these inhibitors were highly selective. Further analogue generation was subsequently performed to identify and optimise lead compounds and generate key structure-activity data with a view of selecting a candidate for in vivo studies.