This study explores how vibe coding, a natural-language approach to building software, allows an educator without programming expertise to create a customized tool for process-oriented English writing. Using Google Antigravity, the researcher developed a web application that guides elementary EFL learners through five writing stages, giving hint-based AI feedback that prompts thinking rather than supplying answers. The application was built through 18 natural-language prompts, the most detailed of which specified pedagogical intent rather than code, and was implemented with 24 fifth-grade Korean learners. Students’ texts grew significantly longer during revising, where the greatest textual change also occurred, and errors rose as texts expanded before falling again during editing. Learners found the feedback accessible and the staged structure helpful, and the classroom teacher could shift attention from giving feedback to managing the lesson, although copying of feedback and inconsistent AI advice emerged as concerns. The study illustrates both the promise and the practical limits of vibe-coded tools for young EFL writers.

1. INTRODUCTION
2. LITERATURE REVIEW
    2.1. Process-Oriented Writing in Elementary EFL Contexts
    2.2. AI-Assisted English Writing Education
    2.3. Vibe Coding and Its Application in English Language Education
3. METHODOLOGY
    3.1. Research Design
    3.2. Research Context and Participants
    3.3. The Web Application
    3.4. Vibe Coding Development Process
    3.5. Implementation Procedure
    3.6. Data Collection
    3.7. Data Analysis
4. RESULTS AND DISCUSSION
    4.1. Developing the Web Application Through Vibe Coding (RQ1)
    4.2. Changes in Students’ Writing Across the Process Stages (RQ2)
    4.3. Affordances and Limitations as Perceived by the Students and theTeacher (RQ3)
5. CONCLUSION
REFERENCES

• Kyuhee Jo(Associate Professor, Department of English Language Education, Gyeongin National University of Education; 155, Sammak-ro, Manan-gu, Anyang-si, Gyeonggi-do, Korea)