The ability to treat osteochondral defects is a major clinical need. Existing polymer systems cannot address the simultaneous requirements of regenerating bone and cartilage tissues together. The challenge still lies on how to improve the integration of newly formed tissue with the surrounding tissues and the cartilage-bone interface. This study investigated the potential use of different silk fibroin scaffolds: mulberry (Bombyx mori) and non-mulberry (Antheraea mylitta) for osteochondral regeneration in vitro and in vivo. After 4 to 8 weeks of in vitro culture in chondro- or osteo-inductive media, non-mulberry constructs pre-seeded with human bone marrow stromal cells exhibited prominent areas of the neo tissue containing chondrocyte-like cells, whereas mulberry constructs pre-seeded with human bone marrow stromal cells formed bone-like nodules. In vivo investigation demonstrated neo-osteochondral tissue formed on cell-free multi-layer silk scaffolds absorbed with transforming growth factor beta 3 or recombinant human bone morphogenetic protein-2. Good bio-integration was observed between native and neo-tissue within the osteochondrol defect in patellar grooves of Wistar rats. The in vivo neo-matrix formed comprised of a mixture of collagen and glycosaminoglycans except in mulberry silk without growth factors, where a predominantly collagenous matrix was observed. Immunohistochemical assay showed stronger staining of type I and type II collagen in the constructs of mulberry and non-mulberry scaffolds with growth factors. The study opens up a new avenue of using inter-species silk fibroin blended or multi-layered scaffolds of a combination of mulberry and non-mulberry origin for the regeneration of osteochondral defects.