issacw228/student-llm-wiki

Student LLM Wiki is a project template that turns your course PDF slides into a structured personal wiki without you writing a single note by hand. You drop your slide decks into a folder called raw/, point an AI coding tool at the project, and type a short command. The AI reads the PDFs, extracts the concepts, and writes organized notes into a wiki/ folder that you then browse in Obsidian, a free desktop and mobile note app. The project works with several AI tools, so you pick whichever one you already use: Claude Code (either the command-line version or the web version at claude.ai/code), Cursor, Trae, Cowork, or GitHub Copilot. Each tool reads a configuration file that the project ships with, so there is no manual setup. You clone the repository, open the folder in Obsidian, install one community plugin called Dataview, and you are ready to import slides. Once you have notes in the wiki, there are four built-in commands you can run from the AI chat panel. The ingest command imports a PDF and generates concept pages, a course overview, and a source summary. The lint command scans for broken links, orphan pages, and stale concepts. The review command runs a Feynman-style quiz where the AI asks you questions and you answer in plain English. The exam-prep command generates practice questions based on the concepts you answered weakly during review. You can also type plain English instead of the exact commands. The wiki is organized into four subfolders: concepts for individual topic pages, courses for per-course overviews, sources for slide summaries, and exam-prep for generated questions. The Home.md file in Obsidian acts as a live dashboard that updates automatically when you add a new course. The same PDF file will never be imported twice because the project tracks files by their content hash. The README is written in both Chinese and English. The project is based on a pattern originally described by Andrej Karpathy and is released under the MIT license.