wangcy7889/bilicourseai

BiliCourseAI is a command-line tool that turns Bilibili course videos into structured study reports. Bilibili is a Chinese video platform widely used for educational content. The tool fetches the video's subtitles and metadata, sends them to an AI model, and produces a local report you can open in a browser. It does not process viewer comments or recommendations, focusing only on the video content itself. The report is laid out with a knowledge tree on the left and study notes on the right. The tree breaks the course into topics and subtopics, which you can fold and unfold to navigate long lectures. Each node on the right side can show formatted notes, formula tables, and automatically captured screenshots from the video. A visual AI model looks at those screenshots to help explain diagrams, equations, and whiteboard content. To use the tool, you install it with pip and then run a sequence of commands: first probe a video to check for subtitles, then outline it to build the knowledge tree, then optionally launch a local web server to interact with the report. Within the browser view, you can expand any node into full notes or break it into finer subtopics. You can also redo individual nodes without starting over from scratch. The tool supports different study approaches depending on the video structure. Long single-part videos get automatically segmented by subtitle timing. Multi-part series can be explored one episode at a time, keeping earlier episodes intact as you add new ones. If the parts are already clearly labeled in Bilibili, you can skip the AI outline step and use the original part list as the starting tree. BiliCourseAI requires Python 3.11 or later and an OpenAI-compatible AI service you configure separately. Subtitle access on Bilibili usually requires a logged-in account, so the tool includes a QR-code login step. The project is described as a prototype suited for personal study and note generation rather than production use.