jbiaojerry/ebook-treasure-chest

The problem

Chinese readers looking for a specific title often find it delisted, region-locked, or split across several reading apps (樊登读书, 微信读书, 京东读书, 喜马拉雅). Hunting it down means jumping between forums, cloud-drive posts and dead links. This repository aggregates download links into one browsable index, organized by tag, with a live search page on GitHub Pages.

What's different

This is not software. It is a curated link list plus a small static search front end — closer to an awesome-list than to a tool. The distinctive choices are breadth of categorization (hundreds of tags, from 文学 down to 咖啡 and 拜占庭) and the claim that each entry carries epub, mobi and azw3 formats. Compared with searching public library or paid subscription services, you get format flexibility but no guarantee of legality, availability or quality; compared with generic book-finding forums, the index is cleaner and searchable but no more accountable for what the links point to.

Who it's for

Readers who already know the title they want and mainly need a fast lookup across many tags, especially on e-ink devices that want epub or azw3. It is not for anyone who needs licensed, citable, or long-term-archived copies — the links are third-party and can rot. People wanting a catalog of their own library, metadata management, or synchronization across devices should use Calibre or a proper reading service instead. Developers looking for a code library will find nothing here.

Cost of adoption

The README documents no installation, runtime, or dependencies; the practical use is opening the GitHub Pages search or grepping the markdown category files. Risks matter more than setup: the license is not declared (the README badge points at a license but the repository metadata says none), so redistribution terms are unclear; link rot and removal are likely; and downloading copyrighted books from third-party hosts may be illegal in your jurisdiction. The repository also accepts donations, which is a maintenance signal but not a quality guarantee.

Verdict

The project is roughly three years old and has accumulated over 22,000 stars, but its trending appearance is recent and brief — about 220 stars per day across two board days in October 2026, with no Hacker News discussion. That pattern looks like a burst of attention on an established link collection rather than a newly built tool. Treat it as a convenience index, not as infrastructure: useful when a search hits, useless when a link dies, and carrying legal and licensing ambiguity you should weigh yourself.