VR technology allows learners to access simulated, immersive and interactive virtual environments to perform authentic learning activities. In particular, VR has emerged as a valuable tool for L2 learning. However, VR research has tended to pay more attention to desktop-based VR than to VR via mobile-rendered HMDs, leaving the potentials of VR through mobile-rendered HMDs yet to be investigated. Therefore, this study fills the gap by using a commercial VR app to examine the effect of VR via mobile-rendered HMDs on EFL learners' vocabulary learning. Forty-nine seventh graders in Taiwan were recruited from two intact classes and assigned to either an experimental (VR players) or control (video watchers) group. The VR players interacted with Mondly VR app using mobile-rendered HMDs and took part in conversations with virtual characters. The video watchers watched the walkthrough video signal of the VR player's app via a personal computer. Vocabulary tests, a perception questionnaire, and interviews were used to evaluate the participants' vocabulary learning. The results showed that the VR players' vocabulary learning and retention was significantly higher than the video watchers'. The majority of the VR players felt that VR-mediated vocabulary learning was motivating and beneficial. The VR app contextualized vocabulary learning by providing virtual environments with multimodal support and enhanced learner engagement through real-time interactivity and feedback. The video watchers' feedback revealed mixed feelings. Some felt that the walkthrough video facilitated vocabulary learning by providing word meaning and use in context. Others reported it lacked interactivity and their attention was easily distracted.