Background. Exercise therapy could potentially modify metabolic risk factors and brain physiology alongside improving function post stroke. Objective. To explore the short-term metabolic, brain, cognitive, and functional effects of exercise following stroke. Methods. A total of 40 participants (>50 years, >6 months post stroke, independently mobile) were recruited to a single-blind, parallel, randomized controlled trial of community-based exercise (19 weeks, 3 times/wk, “exercise” group) or stretching (“control” group). Primary outcome measures were glucose control and cerebral blood flow. Secondary outcome measures were cardiorespiratory fitness, blood pressure, lipid profile, body composition, cerebral tissue atrophy and regional brain metabolism, and physical and cognitive function. Results. Exercise did not change glucose control (homeostasis model assessment 1·5 ± 0·8 to 1·5 ± 0·7 vs 1·6 ± 0·8 to 1·7 ± 0·7, P = .97; CI = −0·5 to 0·49). Medial temporal lobe tissue blood flow increased with exercise (38 ± 8 to 42 ± 10 mL/100 g/min; P < .05; CI = 9.0 to 0.1) without any change in gray matter tissue volume. There was no change in medial temporal lobe tissue blood flow in the control group (41 ± 8 to 40 ± 7 mL/100 g/min; P = .13; CI = −3.6 to 6.7) but significant gray matter atrophy. Cardiorespiratory fitness, diastolic blood pressure, high-density lipoprotein cholesterol, physical function, and cognition also improved with exercise. Conclusion. Exercise therapy improves short-term metabolic, brain, physical, and cognitive function, without changes in glucose control following stroke. The long-term impact of exercise on stroke recurrence, cardiovascular health, and disability should now be explored.