The inadequacy of conventional synthetic grafts has led to efforts to construct a superior vascular graft. In vivo tissue engineering is one approach to this problem that has been investigated for half a century and enables the construction of autogenous vascular prostheses. Three types of in vivo engineering are explored: remodelling of implanted scaffolds, fibrocollagenous tubes, and the artificial artery generated in the peritoneal cavity. Scaffolds designed to be remodelled may be synthetic or biological and have been remodelled in animal models to form vasoactive neoarteries with arterial morphology. The differences in vascular remodelling ability, particularly spontaneous endothelialisation, between animal models and humans may impair the effectiveness of this approach in the clinic. Fibrocollagenous tubes such as the Sparks' Mandril have demonstrated poor performance in the clinic and are prone to aneurysm formation. The artificial artery generated in the peritoneal cavity is a novel addition to the ranks of in vivo engineered vascular prostheses and combines many of the best features of scaffolds designed to be remodelled and fibrocollagenous tubes. However, understanding and manipulating the vascular remodelling process will be the key to producing the ideal arterial prosthesis.