SJTU AI student building practical tools for knowledge workflows, browser organization, and AI productivity.
Currently maintaining Fries, an open-source desktop console for AI quota analytics, multi-account routing, timeline logs, and token insights.
Current work in progress:
- GLM OCR: upgrading it into a configurable OCR toolbox with customizable backends
- Haonote: preparing an OCR-first note workflow that can later extend to local videos and more platforms
- OpenClaw: continuing the unfinished supervision and automation workflows
- English Vocabulary App: a desktop-first vocabulary app for Gaokao, CET-4/6, and postgraduate entrance exam preparation
- League of Legends Multimodal AI Commentary: a long-term project for professional match understanding, analysis, and AI casting
| Project | Focus |
|---|---|
| Fries | Desktop console for AI subscription windows, account routing, timeline logs, token analytics, and multi-platform releases |
| GLM OCR | PDF/PPT/Image to Markdown OCR toolkit based on Zhipu GLM-OCR |
| PDF 4-in-1 | 4-in-1 print layout converter for lecture handouts and slide decks |
| VS Code Frosted Glass | Frosted-glass transparency setup and maintenance workflow for VS Code on Windows |
| Bilibili Favorites | Export Bilibili favorites, classify them with ChatGPT/Claude/Gemini or other frontier LLMs, then sync the resulting folder structure back |
| Bilibili Follow | Export Bilibili follows, classify them with ChatGPT/Claude/Gemini or other frontier LLMs, then sync the resulting groups back |
| Tab Organize | Export Edge/Chrome tabs, classify them with ChatGPT/Claude/Gemini or other frontier LLMs, then import the grouped tabs back into the browser |
| Bookmark Organize | Export browser bookmarks and use frontier LLMs for merge, deduplication, link checking, and category planning before writing the cleaned structure back |
- Knowledge workflow and OCR
- Browser and content organization
- AI-assisted personal tooling
- Desktop productivity apps