TiClearn

TiCLearn is an AI-powered study platform for university and HND students in Cameroon. It helps students turn their actual course materials into meaningful practice, understand their mistakes, and identify what they should study next. During the Meta Global AI Developer Hackathon, we expanded TiCLearn into a more connected learning experience. We built visual course-page analysis that interprets diagrams, charts, and tables in lecture materials; grounded question generation with answer keys, marking rubrics, concept tags, and source-page references; and targeted retries that explain misconceptions and generate follow-up exercises based on what a student got wrong. We also improved the tutor experience with interactive quizzes inside chat, guided quiz setup, one-question-at-a-time practice, immediate marking, mark schemes, saved attempts, weak-area follow-ups, and more reliable fallback paths when AI generation fails. The application uses Muse Spark 1.3 through Meta Model API for its core study reasoning and generation, and Muse Spark Open through Baseten for selective visual analysis of course pages. We also used Muse Code through OpenCode as part of our development workflow. Our goal is to make AI useful beyond answering questions: helping students move from course material to demonstrated understanding and more purposeful study.