Knee osteoarthritis affects 40% of people during their lifetime, significantly impacting societies worldwide. Its molecular pathogenesis remains poorly understood and variable clinical phenotypes suggest it may be more than one disease. We established Synovial fluid To detect Endotypes by Unbiased Proteomics in OA (STEpUP OA) to search for molecular endotypes in knee OA synovial fluid, and to reveal key pathobiological pathways across 1361 individuals with knee OA. Using unsupervised clustering, a single cluster representing a biological continuum is observed, primarily driven by "Epithelial Mesenchymal Transition". Distinct molecular endotypes are not detected. "Angiogenesis", "Complement" and "Coagulation" are enriched for after stratification by clinical phenotype (obesity status, biological sex). Complement and coagulation are associated with the inflammatory marker, C-reactive protein. Associations with patient-reported knee pain are weaker. These findings support knee OA as a biological continuum, identify common and phenotype-enriched targetable pathways, and a rationale for stratification in clinical trial design.