OBJECTIVE: To investigate the STRoke Interactive Virtual thErapy (STRIVE) intervention on upper-extremity clinical outcomes in community-dwelling stroke survivors. DESIGN: Assessor-blinded randomized controlled trial. SETTING: Study screening and testing was conducted in a university clinic. Participants completed the virtual therapy (VT) intervention in a community-based stroke support group setting. PARTICIPANTS: Of 124 stroke survivors initially assessed, 60 participants were recruited (time poststroke, 13.4±8.9 y). Participants were allocated to either VT or control group using a block randomization design and were stratified by sex. INTERVENTIONS: Participants were randomized to receive 8 weeks of VT or usual care. The intervention consisted of approximately 45 minutes of twice weekly VT training on the Jintronix Rehabilitation System. MAIN OUTCOME MEASURES: Between-group differences in the Fugl-Meyer Upper Extremity scale and Action Research Arm Test score were joint primary outcomes in this study. RESULTS: Significant between-group differences for the Fugl-Meyer Upper Extremity scale were seen at the end of the intervention (F=5.37, P=.02, d=0.41). No significant differences were observed with the Action Research Arm Test. No adverse events were reported. CONCLUSIONS: We demonstrated clinically meaningful improvements in gross upper extremity motor function and use of the affected arm after a VT intervention delivered via a community-based stroke support group setting. This data adds to the contexts in which VT can be used to improve upper limb function. Use of VT in community-based rehabilitation in chronic stroke recovery is supported.