Stroke is a leading cause of adult disability. Over 7000 New Zealanders experience a stroke every year. Three months after stroke 20% of people remain wheelchair bound and many of those who can walk do so with inefficient, unstable hemiparetic gait patterns. Almost all patients with limited mobility identify independent mobility as their highest priority. Robot-assisted gait therapy at the acute/subacute stage improves the odds of non-independent patients becoming independent walkers. Furthermore, over-ground gait training is much more promising than traditional stationary treadmill training. Therefore, we plan to undertake the first study of automated over-ground gait training for acute stroke patients who cannot walk independently. Here we will evaluate the feasibility of this approach using a novel gait trainer we have developed. This includes evaluating the safety, usability and acceptance of the intervention as well as calculating an effect size and sustainable recruitment rate for the full study randomised controlled trial.