Objective: The evaluation of a community-based exercise and education scheme for stroke survivors. Design: A single blind parallel group randomized controlled trial. Setting: Leisure and community centres in the south-west of England. Subjects: Stroke survivors (median (IQR) time post stroke 10.3 (5.4-17.1) months). 243 participants were randomized to standard care (124) or the intervention (119). Intervention: Exercise and education schemes held twice weekly for eight weeks, facilitated by volunteers and qualified exercise instructors (supported by a physiotherapist), each with nine participants plus carers or family members. Method: Participants were assessed by a blinded independent assessor at two weeks before the start of the scheme, nine weeks and six months. One-year follow-up was by postal assessment. Main measures: Primary outcomes: Subjective Index of Physical and Social Outcome (SIPSO); Frenchay Activities Index; Rivermead Mobility Index. NHS, social care and personal costs. Secondary outcomes included WHOQoL-Bref. Analysis: Intention-to-treat basis, using non-parametric analysis to investigate change from baseline. Economic costs were compared in a cost-consequences analysis. Results: There were significant between-group changes in SIPSO physical at nine weeks (median (95% confidence interval (CI)), 1 (0, 2): P=0.022) and at one year (0 (-1, 2): P=0.024). (WHOQol-Bref psychological (6.2 (-0.1, 9.1): P=0.011) at six months. Mean cost per patient was higher in the intervention group. The difference, excluding inpatient care, was 296 pound (95% CI: -321 pound to 913) pound. Conclusion: The community scheme for stroke survivors was a low-cost intervention successful in improving physical integration, maintained at one year, when compared with standard care.