Catching the Vibe: Examining Vibe Coding on Human-AI Synergy,in Web Application Development

Generative artificial intelligence (AI) may change how language learning materials are designed, particularly for teachers and students who face technical barriers to developing interactive web resources. This qualitative case study examined how 40 undergraduate students used AI-supported vibe coding to create web applications over a four-week instructional session. Data included web application reports, peer evaluations and student reflections. Class-level analyses examined the language domains students targeted in their interactive applications, whereas individual case-level analyses focused on two students, Sophia and Colin, whose applications received strong peer evaluations. Findings revealed that students most often preferred applications for vocabulary learning, suggesting a tendency to use vibe coding for discrete language skills that could be easily developed into interactive tasks. The focal cases showed a workflow process broadly aligned with the DREAM framework (Xin, 2024). Unlike prior applications of the framework, however, students often drew on their own learning histories to identify instructional needs. This highlights the role of personal experience in developing AI-assisted materials. Overall, students had a positive experience with the process but they also reported feeling frustrated and uncertain when AI-generated outputs failed to function as intended. Peer evaluations were favorable, especially when the applications were perceived as engaging, usable, and responsive to learners’ proficiency levels. These findings suggest that AI-supported vibe coding can expand opportunities for language learning materials development.