Stroke and other neurological pathologies are an increasing cause of hand impairment, involving expensive rehabilitative therapies. In this scenario, robotics applied to hand rehabilitation and assistance appears particularly promising in order to lower therapy costs and boost its efficacy. This work shows a recently conceived hand exoskeleton, from the design and realization to its preliminary evaluation. A control strategy based on surface electromyography (sEMG) signals is integrated: preliminary tests performed on healthy subjects show the validity of this choice. The testing protocol, applied on healthy subjects, demonstrated the robustness of the whole system, both in terms of mimicking a physiological distribution of finger forces across subjects, and of realizing an effective control strategy based on the user's intention.