<![CDATA[The integration of technology into language learning has created new opportunities to enhance effectiveness and learner engagement. However, existing grammar learning applications often lack sufficient contextual scaffolding, provide minimal adaptive feedback, and rely on chatbots that are not grounded in curated learning sources, limiting their effectiveness for structured language acquisition. This study focuses on developing a mobile-assisted English grammar-learning application using the Formula 33 approach and an RAG-based chatbot. The research uses a prototyping methodology, enabling iterative development and continuous improvement informed by user feedback. The application was built using Android Studio (Kotlin), integrated with Firebase Realtime Database, and features a Retrieval Augmented Generation (RAG)-based chatbot that uses the Formula 33 book as its main self-learning source. The implementation results show that all main features, including formula learning, practice exercises, quizzes, and personalized feedback, are present and function effectively. System evaluation using Black Box Testing confirmed functional reliability, while the System Usability Scale (SUS) test yielded an average score of 83.125, categorized as “Excellent” and “Acceptable” for usability. These findings demonstrate that the Formula 33 method, implemented as a pedagogical framework integrated with RAG technology based on curated textbook sources and evaluated through SUS standards, effectively supports users in progressively and contextually understanding English grammar structures. The proposed model offers strong potential as a replicable framework for the development of future technology-assisted language learning applications.]]>
