Pengembangan Website Edukasi dengan Integrasi API ChatGPT sebagai Asisten Virtual Cerdas Menggunakan Django (Studi Kasus Siswa MTs Negeri 7 Klaten)

The rapid advancement of artificial intelligence has significantly influenced the educational sector, particularly in facilitating web-based self-directed learning. Observations at MTs Negeri 7 Klaten revealed that many students encounter challenges in comprehending learning materials independently, as they tend to rely heavily on teacher-centered explanations and have limited access to interactive learning resources. These findings highlight the necessity for an educational system capable of providing adaptive and responsive learning support. This research focuses on the development of a Django-based educational website integrated with the ChatGPT API to function as a virtual learning assistant. The system was developed using the Waterfall model, encompassing the stages of requirements analysis, system design, implementation, testing, and maintenance. ChatGPT was utilized to deliver interactive learning features, including question-and-answer services and conceptual explanations to support independent learning. System evaluation was conducted through Black Box testing to assess functionality and the System Usability Scale (SUS) to measure usability. The Black Box testing results confirm that all system features functioned as intended, while the SUS assessment yielded an average score of 62,25, categorized as Marginal, indicating that the system is usable but requires further refinement to enhance overall usability.