This study investigates informal English as a Foreign Language (EFL) learning in immersive virtual reality (VR) environments, focusing on play-based interactions within a community of DVVerb users in VRChat. The research aims to examine how language development, identity formation, and peer interaction emerge in unstructured virtual spaces. Employing a qualitative virtual ethnography approach, data were collected through 90 hours of participatory observation, twelve semi-structured voice interviews, and multimodal artifact analysis. Findings indicate that VRChat serves as an affinity space that promotes meaningful, self-directed English practice beyond formal instruction. Participants reported improvements in speaking fluency, pragmatic competence, and sociocultural awareness through game-based tasks, role-play with avatars, and peer scaffolding. Avatars and spatial cues reduced speaking anxiety and encouraged risk-taking, while regular social rituals, such as storytelling nights and team quests, provided continuous language opportunities and peer feedback. Challenges included ineffective language use, potential disruptions, technological limitations, and reliance on learner motivation. The study concludes that VR environments offer rich, immersive contexts for informal EFL learning, supporting autonomous, socially mediated, and experiential development. These findings have practical and theoretical implications for digital language education, demonstrating that virtual “digital tribes” can foster community-based learning and complement traditional classroom instruction.