Compile a field into a structured knowledge repo where every entry is tagged with an evidence level and source. Chinese first, but it can be replicated into legal, health insurance, or immigration guides, exam past papers, and subject knowledge graphs in any language. Open source reliably earns stars; the path to revenue hasn't appeared yet.
- Status
- Open (stars keep coming, derivative repos are growing; no revenue sample at all)
- Why now, and how long the window lasts
- AI has cut the time to systematically compile a field into a book or a map from years to weeks, while readers still trust sourced, graded, human-curated material (inference). Chinese internet lacks trustworthy structured resources on daily life and education, so GitHub has become a substitute publishing channel, and spread happens via WeChat and Xiaohongshu (none of the five X circles mentioned it). Don't chase events in this category. There are plenty of gaps: after cn-math, chemistry, history, and biology are all empty; for legal, health-insurance, and immigration guides, apart from Brazil, other languages are nearly empty.
- Confidence
- Medium (the star pattern shows 1 giants plus 6+ similar repos at 2,000–6,000 stars); the revenue side is low. What would overturn this: HowToLiveBetter's 11-15-day star gain falls below 7 500, and no new repos of the same type appear. What would upgrade it: the first product that earns money from this kind of repo (an app, mini-program, or print edition), at which point I'd switch to tracking.
Evidence
Grouped by store. Each group starts with the charts that can be graphed (real curves, data dates below the chart), then the scanner's original notes and source links.
GitHub5 evidence · 13 products with curves
5 more products' curves
- GitHub (full-site snapshot, 11-02 verified): eternity4719/HowToLiveBetter "High-Value Life Guide" (longevity and disease prevention, first aid, saving money, legal red lines, medical and social insurance…; each entry states cost, benefit, evidence grade, and source, citing only journals and official documents). Created 09-07, 72,286 stars (report said 11-01 was 71,880); +7 in 3,349 days, +7 in the previous 3,217 days, and it still hasn't slowed down 56 days after creation; the 09-28 → 10-09 window gained 11 30,709; forks 8.8%, never posted on HN.
- GitHub (derivatives, 11-02 verified): single-page reading edition cdyforever/how-to-live-better 13,984 stars; overseas edition BeckyGlobal/PaidToStudyAbroad ("High-Value Study Abroad", 71 entries, citing only official documents) created 10-15, 1,342 stars; mini-program edition HuiTurn/HowToLiveBetter 219 stars. One content repo grew three kinds of derivatives (a reader, a mini-program, and an overseas edition) within 8 weeks.
- GitHub (Chinese same-type, 11-02 verified): xiaoyang111222/cn-math-taxonomy (knowledge graphs for primary/secondary math, English, and junior-high physics; 1,644 knowledge points, 6,723 questions), created 10-18, 5,840 stars, +2,201 in 7 days; designfuchen/word-atlas (IELTS vocabulary illustrations + FSRS review), created 10-25, 3,695 stars; anhao/ai-pose-gallery-zh (pose reference gallery), created 10-27, 5,115 stars; ERRRC/xingcezhenti (2016–2026 national and provincial civil service Xingce questions), 2,147 stars, and the same author's ERRRC/kaogongzhentizhengliu, reorganized by topic, 1,007 stars.
- GitHub (Chinese same-type, continued, 11-02 verified): DeanChensj/jiangshan-map (interactive China historical map), 2,725 stars, +11-02 in a single day 668; RadishXN/run-abroad (fill in 10 details, filter the 71 visa / immigration routes), 1,286 stars, +7 in 760 days, and only +7 in the previous 16 days, so it was suddenly discovered.
- GitHub (non-Chinese same-type, 11-02 verified): ctresb/leis (all of Brazilian law), created 10-15, 2,224 stars; lexifina/global-law (each country's law as a git history), created 10-28, 1,555 stars.
Indie product revenue (TrustMRR)1 evidence · 2 products with curves
- indie(11-02 searching visa / immigration / exam / evidence and similar terms): no product monetizing this kind of structured knowledge repository turns up. The ones with visible revenue are a different breed: single-country visa service sites (Get Brazil Visa $3,739/30 days) and exam question bank sites (CCAT Tests $207). None of them are open-source content.
Already doing it
HowToLiveBetter and its three derivatives, cn-math-taxonomy, word-atlas, xingcezhenti, run-abroad, jiangshan-map, ai-pose-gallery-zh, leis, and global-law, are all free. Book-style projects like bojieli/ai-infra-book (7,220) and LAMDA-NeSy/Research-Starter-Kit (9,416) follow this path too. The report lists Renhuai123/ziwei-doushu (a Zi Wei Dou Shu chart engine, 13,018) as a contemporary, but it was created 05-01 earlier, and it's a code engine, so it's only a loose match.
The hardest step
Accuracy and ongoing maintenance. Legal, health-insurance, and exam policies change every year. Once people see it as "AI-generated encyclopedia," reputation collapses. Second, monetization: this category currently earns stars, not money.
Who it's for
An individual who knows a field, has the editorial patience, and can work in a non-English language. Possible ways to earn: a paid app or mini-program edition, print editions, country-by-country update subscriptions. All of these are inference; there's no sample yet.
Should every product builder know this?
No. But the fact that a pure content repo can reach tens of thousands of stars and still grow derivatives is worth knowing for anyone working on developer content and community.
Change log
Original: Scanner opportunity library #054 (updated daily, 06:00); this page was generated on the 2026-11-08 edition.