This paper presents a new eHealth platform incorporating humanoid robots to support an emerging multidimensional care approach for the treatment of diabetes. The architecture of the platform extends the Internet of Things to a Web-centric paradigm through utilizing existing Web standards to access and control objects of the physical layer. This incorporates capillary networks, each of which encompasses a set of medical sensors linked wirelessly to a humanoid robot linked (via the Internet) to a Web-centric disease management hub. This provides a set of services for both patients and their caregivers that support the full continuum of the multidimensional care approach of diabetes. The platform’s software architecture pattern enables the development of various applications without knowing low-level details of the platform. This is achieved through unifying the access interface and mechanism of handling service requests through a layered approach based on object virtualization and automatic service delivery. A fully functional prototype is developed, and its end-to-end functionality and acceptability are tested successfully through a clinician-led pilot study, providing evidence that both patients and caregivers are receptive to the introduction of the proposed platform.