The trend of self-directed learning continues to rise, yet students still face challenges such as limited time, difficulty structuring learning materials, and a lack of systematic learning pathways. This study aims to design and develop a web-based AI Course Generator application by integrating the Gemini API to generate automated learning materials, the YouTube API to recommend educational videos, and an interactive chatbot to provide personalized learning assistance.The development method uses the Rational Unified Process (RUP), consisting of four phases: inception, elaboration, construction, and transition, supported by Unified Modeling Language (UML) modeling. The implementation results show that the application can automatically generate course outlines and learning materials, recommend learning videos, and assist users through the chatbot.Testing was performed using black box testing and usability testing with the USE Questionnaire, yielding a user satisfaction score of 82.09% (strongly agree category). This study demonstrates the effectiveness of integrating the Gemini API and YouTube API in building an adaptive and efficient self-learning application that supports the Merdeka Belajar concept.
Copyrights © 2026