I need a macOS desktop application that automatically creates multiple-choice exams from an academic PDF using AI. I want to: • Upload PDFs and have the system analyze their content without requiring me to input a set of questions. • Choose, before each generation, how many questions the exam will have and how many answer choices each question will display. • When answering, have the application indicate whether I answered correctly or incorrectly and provide the justification, citing the exact article, section, or page of the syllabus that supports the correct answer. Main technical requirements: • Native execution on macOS (an installable .app or equivalent), without a mandatory connection to external servers once installed, except to call the AI API of your choice. • Simple interface: PDF selector, numeric fields for questions and answers, "Generate Exam" button, and an interactive quiz view with immediate feedback. • Basic persistence of results to review previous attempts. • IT CAN ALSO BE ON ANDROID I welcome suggestions regarding the AI model and PDF library you consider suitable for macOS (for example, Core ML, Python + PyObjC, Swift + PDFKit, etc.). I would appreciate a small demo or functional prototype that demonstrates the generation of questions from a sample chapter.