LearnMore

As a master's student studying abroad in English, which is not my first language, I need to understand unfamiliar ideas, put them in my own words and connect them across lectures. I built LecNote, a macOS lecture-PDF app, around this learning process. When I get stuck on a sentence, formula or diagram, I can select it and ask AI with the surrounding lecture context, or attach a picture to my question. Answers start short, with points I can expand when I need more depth. The explanation language can differ from the interface language. I can type notes beside the slide and insert a same-sized blank page after it when I need room for a longer explanation or derivation. During this hackathon, I built the one-click pipeline that creates a knowledge structure and formula sheet from a course's slides. It groups concepts across lectures and links them back to source pages. Formulas are linked to those concepts, with symbol explanations, conditions and source references. I also added bilingual knowledge-map views with revealable key points, collapsible formula groups and a course-wide page of unresolved questions that I can answer and mark resolved. Muse Spark 1.3 reads the slide pages; Muse Wonder plans, drafts and reviews the generated material. Code checks structure, page references, quotation matches and formula rendering, and detected issues feed repair tasks. Results appear as a preview before I choose what to adopt. A recorded Meta run on three original demo lectures produced both outputs for $1.57; all 46 saved source quotations passed text-matching checks against their cited pages. New work includes the generation pipeline, Meta integration and study-workflow improvements above. The original reader, basic notes, blank-page insertion and original Ask AI existed before the event. The standalone repo isolates the generation core and includes a runnable Meta CLI for reading, planning and repair; the full desktop workflow is shown in the demo.