ruyiJS: a new tool from a decade-old reverse-engineering training brand, reposted 404 times, reached 9 stars in 759 days
What it is It packages the real Firefox JS engine and network stack into a headless browser callable from Python, used for running website encryption signatures, getting past anti-bot systems and captchas.
Why it blew up The author runs a reverse-engineering training brand with 70k students over ten years, and comes with WeChat Official Account and group-chat followers built in. The linux.do launch post's title named a dozen-plus domestic and foreign shields outright. After the original repo was 404 and rebuilt, another wave hit with people forking it to keep a copy.
CHow you can ride this Don't build another tool like it. Copy the launch approach: put in the README's first screen which specific problems it solves, write examples for AI to read, launch in vertical Chinese communities with an open-source promo post and reply to every comment. If you want to follow this direction, build a compliant version for the English-speaking world, or a version that supports older Linux.
- Data (updated daily) · Repo created 2026-10-28 · 889 stars · 7 days +536
GitHub · stars Data 2026-10-29 → 2026-11-08 (11 days, 11 observations). Points before 10-28 come from deeply / archive backfill, not every day has one
ruyiJS extracts the real Firefox JS engine and network stack into a "browser with no window" you can pip install, specifically for running website encryption signatures, bypassing anti-bot systems and captchas. The repo was created in Oct 28 (this is the second repo, the first was taken down by 404), and by Nov it had 6 stars and 759 318 forks. The stars didn't come from the English-speaking world; they came from the Chinese JS reverse-engineering community. Below: what it is, where the stars came from, who it's competing with, and what you can learn from it if you want to build a project that gets similar stars.
A linux.do post, and forty minutes later someone recognized the author
On the evening of Oct 2026 at 26:10 51 (Beijing time), linux.do user tisoz posted with a very long title:
[FastAI · Open-source promo] An anti-bot browser built for AI — one-click bypass of Akamai, DataDome verification, Alibaba 234 bx-ua, JD h5st, Tencent Tianyu, NetEase Yidun, Shumei, Geetest, Boss Zhipin, Douyin, Xiaohongshu, Google Search and other web shields……
A minute later the first reply: "Here it comes, looks like exactly what I need." Then "Insanely strong, bookmarked first" and "Star + bookmark right away, must be useful." Forty minutes later someone recognized who made it: "Oh, isn't this ruyiBrowser from Ruyi? I've been following their WeChat Official Account." Then "Amazing, a Ruyi product." The 21th reply was just one line: "Won't say much, when does it start charging?"
By Nov 6, the post had 2,761 views, 291 likes and 82 replies; the GitHub link in the post was clicked 350 times (linux.do, 11-06).
"Ruyi" is not a new name in the Chinese reverse-engineering scene. It is run by the same people as "Xiaojianbang Education"; its website says "Ten years of security reverse-engineering training, serving 7w+ students cumulatively," with courses on crawlers, JS reverse engineering, Android/iOS reverse engineering, and captcha recognition. It had already released the Chromium fingerprint browser ruyiBrowser, the Firefox automation framework RuyiPage, and the trace tool ruyiTrace. ruyiJS is the latest in this line of products.
What it actually does
The daily problem for JS reverse engineers: a site computes an encrypted parameter with obfuscated JS (JD's h5st, Alibaba's bx-ua, Akamai's sensor_data), and without that parameter requests get blocked. Traditional approaches are two: one is "extracting code," pulling the JS out and reading it by hand; the other is "environment patching," using simulated environments like jsdom to pretend to be a browser and patching whatever is missing. The English-speaking habit is to launch a whole Chrome or Firefox and drive it with Playwright.
ruyiJS sits in between:
- The engine is real. It uses the SpiderMonkey engine from Firefox 159 and the Necko network stack. The author said in the post, "It's the real Firefox… if detected, we patch it right away," so no environment patching is needed.
- No window, no rendering. It's lightweight, so you can run many concurrent instances. The author posted a Bilibili video titled "Akamai's 60-set TLS unbeatable concurrency with ruyijs."
- Call page functions directly from Python. The README example is
context.eval("window.__sign(payload)"), using the site's own signing function as a plain Python function; fingerprint, TLS fingerprint and proxy are configured per Context. - Examples written for AI to read. The package includes many runnable examples for "bypassing a given shield." The author's words: "Just have the AI pip install this library and use it, since it comes with lots of examples."
Where the stars came from
| Date | Cumulative stars | Cumulative forks | What happened |
|---|---|---|---|
| 10-26 | — | — | Launched on linux.do, back when it was the original repo |
| 10-28 | 0 | — | Original repo 404, rebuilt in the afternoon under a new account |
| 10-29 | 95 | 108 | Forks outnumber stars |
| 10-31 | 241 | 213 | Trendshift #5 |
| 11-02 | 400 | 250 | Trendshift #4 |
| 11-03 | 537 | 274 | +137 in a single day, the biggest of the whole run |
| 11-06 | 759 | 318 |
(GitHub stars and forks, daily; Trendshift ranking; 11-06 is the GitHub page figure for that afternoon.)
First, the fans of an established brand. Several replies under the post recognized the author: "Ruyi master! I often read your Official Account articles," "Ruyi master is here, reporting from the Official Account's old fan base." A decade, tens of thousands of students, and one Official Account: the cold-start audience was already there. The README also carries a WeChat group QR code, with the group named "AI browser ruyijs anti-detection discussion group 3"; the first two groups are already full.
Second, it got 404 once, and came back even stronger. On the morning of Oct 28, several people in the thread reported that the GitHub link was dead, asking "Did this get sanctioned?" and "Did they already send a cease-and-desist?" That afternoon the author replied: "Reborn from the dead, the original project was 404, now I've updated the link." The new repo got two commits that same day: trimming the README examples down to just Akamai and CreepJS, and removing an Expedia example. The English README now only names three overseas targets, Akamai, DataDome and CreepJS; all the China-side targets from the launch title (Tianyu, Shumei, Yidun, BOSS, Douyin, Xiaohongshu) are gone, and someone in the thread asked, "Where did the xhs example go?" Why the original repo was taken down wasn't confirmed; a cease-and-desist or a platform report is the most likely cause (inference).
Users of a tool that was deleted once instinctively save a copy first. Trendshift recorded 10-28 with 39 new forks and 0 new stars, and 10-29 with 137 new forks and 3 new stars. The fork-to-star ratio is 0.42, well above what most trending projects show (inference). Spot-checking recent fork accounts, most are real accounts registered for years with dozens of repos, not obviously botted.
Third, several Chinese channels pushing at once. The Bilibili account "Ruyi-AI Browser Dev" posted "Tencent Tianyu invisible/slider captcha instantly beaten with ruyijs" (about 800-900 views); Tencent News carried "Full code for one-click DataDome bypass with ruyijs" (10-29); and YouTube Shorts had an English one, "Bypass any anti-bot check from Python with ruyiJS."
Not showing up: On X, Hacker News, and Reddit, almost nobody talks about it, and no media has covered it.
People are actually using it. PyPI gets 2,385 downloads a week (pypistats, 11-06). For a niche library that only ships a few native platform packages, that's not nothing (mirrors and CI inflate it a bit). Outside contributors are pitching in too: someone added Linux support, someone added macOS support and a batch of DOM performance fixes, someone fixed an HTTP/2 deadlock, and someone built an MCP wrapper on top of it (24 stars).
The money is outside the library. The library itself charges nothing. The post title prefix "【FastAI·开源推广】" and the sponsor logo at the top of the README both belong to FastAIToken, a Claude/GPT API relay. The link in the post got clicked 258 times. In October 30 the license changed to AGPL plus a commercial license, so closed-source use means buying via WeChat. Add WeChat groups and an Official Account that funnel people to courses, and you get a very typical Chinese open-source promo playbook: open source for reach, then pull people into private channels, courses, and the relay.
Who it's competing with
"Anti-detect browsers" is already crowded in the English-speaking world, and the players there are much bigger:
| Project | Launched | Scale (stars) | How it differs from ruyiJS |
|---|---|---|---|
| Camoufox | 2024-07 | 13.1k | Also patches Firefox for anti-detection, but as a full browser, driven by Playwright |
| Lightpanda | 2023-02 | 41.5k | Also pitched as a "headless browser for AI", but written from scratch (Zig + V8), not real Firefox |
| CloakBrowser | 2026-02 | 32.8k | Stealth Chromium that can directly replace Playwright |
| undetected-chromedriver / nodriver / patchright | 2019–2024 | 12.9k / 4.9k / 5.1k | Patches the Chrome or Playwright driver layer, still a full browser |
| curl_cffi | 2022-03 | 6.8k | Only mimics TLS fingerprints, doesn't execute JS |
| sdenv | 2024-03 | 823 | Chinese "environment patching" framework aimed at Ruishu; ruyiJS doesn't need simulated environments |
| PythonMonkey | 2022-09 | 974 | Also embeds SpiderMonkey in Python, but only the engine, with no DOM, network stack, or fingerprint |
| CAPTCHA-solving platforms and commercial anti-bot-bypass APIs | Already existed | — | Paid black boxes; ruyiJS runs locally and is open source |
ruyiJS occupies a very narrow spot: a real engine, no rendering, eval directly in Python. In the Chinese reverse-engineering scene this spot is new; in the English scene, almost nobody knows it exists.
What it still gets wrong
- Picky about platforms. macOS needs Apple silicon on 26 or later, Linux needs glibc 2.39 or later (Ubuntu 24.04 and up), so old servers won't install it.
- It could get taken down at any time. It has already been 404 once, and the China-market examples have been pulled. The most attractive part is exactly the part that can't be published openly.
- The English-speaking world is blank. The README has an English version, but there's been no English-community launch at all.
If it were you
One-line verdict: the stars mostly come from the decade of fans "Ruyi" built up, plus a direction sensitive enough to get a repo deleted. Followers get neither, or don't want it. What's worth learning is how it was launched, not building another one.
Who it's for: people who already have an Official Account, courses, or groups in some vertical tech circle (reverse engineering, scraping, security, a specific framework), and who have a concrete problem insiders deal with every day.
Who it's not for: people who want to ride the buzz by making a "Chinese Camoufox" without any community of their own, or anyone unwilling to take on the risk of takedowns and lawyer letters.
How to do it:
- Put "can this solve my specific problem?" on the first screen of the README. ruyiJS's title lists more than a dozen anti-bot systems by name, so readers instantly know whether it's relevant to them.
- Write the examples for AI. Ship a pile of runnable examples so users can hand the library to a coding agent and get working code back. This is the new selling point for open-source tools in 2026.
- If you want to follow this direction, do the parts it does poorly: a version for the English-speaking world limited to compliant use cases (testing your own sites, collecting public data, fingerprint-detection research), or a version that supports old Linux and fits in Docker. Both gaps look empty right now (inferred, not checked one by one).
How to get stars: for the launch, pick a high-density Chinese tech community (something like linux.do), use the "open-source promo" format they accept, reply to every comment in the thread, and add case studies live. Follow up on Bilibili, the Official Account, and WeChat groups at the same time. New repos don't make GitHub Trending, but third-party charts like Trendshift can give a few days of exposure.
How long the window lasts: the "browser for AI" hype is still rising, probably for the next few months. But what really sets the window is the platforms' stance. If ruyiJS gets deleted again, there's even less room for this direction to operate openly on GitHub.
Signs you've misread it: your launch post gets no more than a few hundred likes in a week and your stars stall in the double digits, or you find that your most compelling examples don't dare go into the README. The first means your circle isn't big enough or your pain point isn't specific enough. The second means you picked a direction that can only be done in private.
Research notes: competitor table, acquisition evidence, full research report
- Competitors
Name Where Launched Scale How it differs from this one Camoufox Open source 2024-07 13.1k★ Camoufox is a full anti-detect Firefox driven through Playwright. ruyiJS doesn't render; it evals the page's signature function directly in Python. Lightpanda Open source 2023-02 41.5k★ Lightpanda is a headless browser written from scratch (Zig + V8). ruyiJS uses real Firefox components, so its fingerprint is real Firefox. CloakBrowser Open source 2026-02 32.8k★ Stealth Chromium that can replace Playwright. Full-browser route. undetected-chromedriver / nodriver / patchright Open source 2019–2024 12.9k / 4.9k / 5.1k★ All patch the Chrome/Playwright driver layer and use a full browser. curl_cffi Open source 2022-03 6.8k★ Only mimics TLS fingerprints, doesn't execute JS sdenv Open source 2024-03 823★ Chinese "environment patching" framework (a modified jsdom) aimed at Ruishu. ruyiJS runs on a real engine and doesn't need simulated environments. PythonMonkey Open source 2022-09 974★ Only the SpiderMonkey engine, with no DOM, network stack, or fingerprint. CAPTCHA platforms / commercial anti-bot-bypass APIs Other Already existed Big Paid black boxes; ruyiJS runs locally and is open source. - Where users come from
- Verified Old brand fans (Ruyi / Xiaojianbang education Official Account, students): linux.do replies say "Ruyi, I often read your Official Account articles" and "Reporting in as an old Official Account fan"; the Xiaojianbang website says ten years and 7k students.
- Verified linux.do open-source promo thread: 2026-10-26 posts, 2,761 views, 291 likes, 82 replies, GitHub link clicked 350 times.
- Verified Bilibili / Tencent News / YouTube Shorts short-form: Bilibili account "如意-AI浏览器开发" posted the Tianyu and Akamai videos (about 800–945 views); Tencent's Penguin account 2026-10-29 posted a DataDome article; one YouTube Short.
- Verified WeChat group private traffic: the README shows the author's WeChat and a QR code for the "anti-detection discussion group 3".
- Verified Fork wave after the delete-and-rebuild: 10-28 original repo 404, and the author replied "Reborn from the dead". Trendshift logs 10-28/29 new 39 and 137 forks, while stars only went up by 0 and 3.
- Verified Trendshift chart: 10-30 #6,10-31 #5,11-02 #4
- Verified Joint promotion with API relay FastAIToken: post prefix "【FastAI·开源推广】", sponsor logo pinned at the top of the README; the link in the post was clicked 258 times.
Why ff-ruyijs (ruyiJS) works
In one line: it extracts the SpiderMonkey engine and Necko network stack from real Firefox and packages them as a "windowless browser" you can pip install into Python, built for JS reverse engineering, running site signatures, and getting past CAPTCHAs and anti-scraping. The author is "Ruyi", a long-running name in Chinese JS reverse-engineering training. The stars come from the author's own fans and Chinese reverse-engineering/scraping communities, not from the English-speaking world.
Repo: github.com/cutetailbud/ff-ruyijs, PyPI: ruyijs, AGPL-3.0 plus a commercial license.
1. Data
| items | value | Source |
|---|---|---|
| Repo created | 2026-10-28 (the first commit is "v0.1.3 initial public source release" from 2026-10-26, which indicates this is a rebuilt repo) | GitHub API |
| Star | Recall card 658 (11-06), live the same day 759; last 7 days +499; single-day peak +137 (11-03) | Data card / GitHub API 2026-11-06 |
| Fork | 318, forks/stars about 0.42 | GitHub API |
| PyPI downloads | Last 7 days 2,385, previous day 131 | pypistats.org 2026-11-06 |
| Release asset downloads | Each release gets dozens to just over a hundred (0.3.0's Windows runtime zip peaked at 115 downloads) | GitHub Releases API |
| Trendshift | 10-30 ranked #6,10-31 #5,11-01 #23,11-02 #4 | trendshift.io/repositories/325142 |
| GitHub Trending | Repo is too new (under about 60 days old) to make trending; trending_days_any=0 | Data card |
| Author account | cutetailbud, registered 2025-05, 18 followers, 5 public repos (the rest are empty repos or forks of firefox): a sock puppet / alt account | GitHub API |
| Committers | charlesval 136 commits, esword 9 commits, Facetomyself 8 commits, plus cursoragent / codex commits | GitHub commits |
2. Why it took off
1. The author is a long-established name in Chinese reverse-engineering training, with a built-in fan base
- Several people under the linux.do launch post recognized the author: "Isn't this the ruyiBrowser from Ruyi? I've been following the Official Account", "Ruyi, I often read your Official Account articles", "Ruyi is here, old Official Account fans reporting in", "Amazing, a Ruyi product" (https://linux.do/t/topic/3078486).
- "Ruyi" and "Xiaojianbang Education" are the same team: their site says "Xiaojianbang and Ruyi have run a security reverse-engineering training program for ten years, serving 7w students in total", with courses on crawlers, JS reverse engineering, Android/iOS reverse engineering, and captcha solving (https://0xshoulderlab.site/). They had already shipped a series of "ruyi" products: the Chromium fingerprint browser ruyiBrowser (Esword618/ruyiBrowser, 57★, whose README says it was "developed from the Ruyi education course"; Esword618 is esword, a contributor to this repo), the Firefox automation framework RuyiPage, and the trace tool ruyiTrace (featured in the Tencent News article "JS Reverse Engineering Finale Plan" from 2026-06-23).
- So ruyiJS didn't start from zero. It's a new product from a training brand that has run for ten years, with tens of thousands of students and a WeChat Official Account.
2. Launched on linux.do, in the "open source promotion" post format
- 2026-10-26 post with a blunt title: "A shield-breaking browser built for AI: one click gets past Akamai and DataDome verification, Alibaba 234 bx-ua, JD h5st, Tencent Tianyu, NetEase Yidun, Shumei, Geetest, Boss Zhipin, Douyin, Xiaohongshu, Google Search and other web shields." 2026-11-06:2,761 views, 291 likes, 82 replies; the GitHub link was clicked 350 times, and the sponsor FastAIToken link 258 times (from the Discourse JSON link_counts).
- Replies were mostly "starred and bookmarked, will be useful someday." Some asked whether it could get past Douyin, Xiaohongshu, Cloudflare, reCAPTCHA and Amazon, and the author answered each one in the thread, live-adding examples for Akamai, Alibaba 234 and DataDome. The audience is JS reverse engineers, freelance scrapers and cross-border e-commerce data collectors, and their pain points are very concrete.
3. The original repo was 404, and a new account "reincarnated" it. Stars and forks actually surged.
- 10-28 morning, several people in the thread reported "the GitHub link is dead, did it get sanctioned?", "did they get a cease-and-desist?", "it's already 404." That afternoon the author replied: "It's reincarnated. The original project was 404, and the link is updated now."
- The rebuilt repo committed "docs: trim README examples to akamai and creepjs only" and "remove the expedia-taap example" the same day. The English README now covers only three foreign targets: Akamai, DataDome and CreepJS. The domestic targets from the launch post (Ruishu, Geetest, Tianyu, Shumei, Yidun, BOSS, Douyin, Xiaohongshu) are gone. Someone asked in the thread, "Where did the xhs example go?" The likely reason for the takedown is fear of domestic risk-control vendors or platforms sending legal letters, but this is unverified.
- Forks are running ahead of stars. Trendshift logged 10-28 with 39 new forks and 0 stars, and 10-29 with 137 new forks and 3 stars. Roughly half the forks came in right after the rebuild, before stars had caught up. A tool that got deleted once makes users's first instinct "fork a copy to keep," which explains why forks are so high, and it doesn't look manufactured. In a sample of the last 12 fork accounts, most were real people with accounts registered 2016–2023 years ago and dozens of repos (caoshuisi, Big1moster, p9s with 272 repos, 16tool registered 2012 years ago, and so on).
- Issue #1 asks, "Is this the original ruyijs project?", which shows some users followed old links all the way here.
4. Content pushed across several channels at once
- Bilibili account "Ruyi-AI Browser Dev" posted "Tencent Tianyu invisible / slider captcha one-click takedown with ruyijs" (around 800–945 plays) and "Akamai, 60 TLS-proof concurrency with ruyijs," among other videos (from the Bilibili search results page).
- Tencent News / Enterprise account: "Complete code for one-click DataDome bypass with ruyijs" (2026-10-29, news.qq.com/rain/a/20261029A005GR00).
- YouTube Shorts: "Bypass any anti-bot check from Python with ruyiJS."
- WeChat Official Account plus WeChat group: the README shows the author's WeChat and a QR code for the "AI browser ruyijs anti-detection discussion group 3" (which suggests the 1 and 2 groups are already full).
- Trendshift listed it. From 10-30 to 11-02 it reached the top 25 on four days (best rank #4), which gave it some exposure beyond the Chinese-speaking audience.
- X: a search for "ruyijs" in the local x-radar corpus returned zero hits in the last 30 days, so almost nobody is talking about it on English X. No results on HN or Reddit either.
5. People are actually using it, and outside contributors have shown up
- 2,385 PyPI downloads per week. For a niche native library that only ships Windows, macOS 26 and glibc 2.39 wheels, that means people are installing it (mirrors and CI inflate the number somewhat).
- Of the 29 issues/PRs, most are PRs from outside contributors: Starktomy added the Linux build, Facetomyself added macOS support and a batch of DOM performance optimizations, and sbbbsbbb fixed an HTTP/2 deadlock. Facetomyself also built ruyi-mcp (24★). The community is putting in real work.
6. Where the money comes from
- It doesn't charge for the library itself. Four lines bring in money: (1) a sponsor, FastAIToken, an API relay for Claude/GPT. Post titles carry the prefix "【FastAI·开源推广】" (FastAI open-source promo), the body says "this project was built with the help of claude1.5x pure-blood opus5.5 ... you can use our relay," the logo sits at the top of the README, and the link in the post has been clicked 258 times. (2) 10-30 switched to AGPL plus a commercial license; closed-source use requires buying a license via WeChat. (3) A WeChat group funnels people into training courses (小肩膀/如意). (4) A WeChat Official Account.
- This is a textbook Chinese open-source promotion playbook: use open source to build attention, then route that audience into private groups, courses and relay services to monetize.
III. Competitors
| Name | Where | Launched | Scale | How it differs from ruyiJS |
|---|---|---|---|---|
| Camoufox (daijro/camoufox) | Open source | 2024-07 | 13.1k★ | Also a Firefox anti-detection modification, but it's a full browser driven through Playwright. ruyiJS doesn't render; it runs only JS and the network stack, and calls page functions directly with eval from Python. |
| Lightpanda | Open source | 2023-02 | 41.5k★ | Also pitched as a "headless browser for AI," but it's written from scratch (in Zig, with a V8 JS engine). ruyiJS uses the real Firefox components, and its selling point is "the fingerprint is real Firefox." |
| CloakBrowser | Open source | 2026-02 | 32.8k★ | A stealth build of Chromium that can directly replace Playwright. It takes the full-Chromium route. |
| undetected-chromedriver / nodriver | Open source | 2019 / 2024 | 12.9k / 4.9k★ | Modifies the Chrome driver layer and still uses full Chrome. |
| patchright | Open source | 2023-11 | 5.1k★ | A patched Playwright. |
| curl_cffi | Open source | 2022-03 | 6.8k★ | Only imitates TLS/HTTP fingerprints and doesn't execute JS. ruyiJS can produce the tokens generated by a site's own JS. |
| sdenv | Open source | 2024-03 | 823★ | A Chinese "environment patching" framework (a modified jsdom) built specifically to get past Ruishu. ruyiJS doesn't simulate an environment; it runs on the real engine directly. |
| PythonMonkey | Open source | 2022-09 | 974★ | Also embeds SpiderMonkey in Python, but ships only the engine, without DOM, network stack or fingerprinting. |
| Scrapling / Botasaurus | Open source | 2024 / 2023 | 98.7k / 5.8k★ | A general-purpose crawling framework that relies on an external browser for anti-bot evasion. |
| browser-use / Steel | Open source plus cloud. | 2024 | 120k / 7.9k★ | An AI agent browser aimed at operating pages, not reverse-engineering signatures. |
| Captcha-solving platforms and commercial shield-bypass APIs (2captcha, various JS reverse-engineering freelancers). | Other | Already existed | Big | Paid, and runs as a black-box API. ruyiJS runs locally and is open source. |
4. Differentiation
In Chinese JS reverse-engineering workflows, when you hit encrypted parameters (h5st, bx-ua, Ruishu cookies, Akamai sensor_data) you either "peel out the code" or "patch the environment" (jsdom, sdenv-style simulated environments). In the English world, you usually launch a full Chrome or Firefox and drive it with Playwright. ruyiJS sits in between: the engine is real, so no environment patching is needed; there's no rendering or window, so it's light and can run many concurrent sessions ("60 TLS concurrency"); and it talks to Python directly, so context.eval("window.__sign(payload)") lets you call the page's own signing function like any normal Python function. Fingerprint, TLS fingerprint (Firefox 100–159) and proxy are all configured per Context. On top of that, it ships with runnable examples for each shield, so a user can hand the library to an AI agent, which can adapt working code from the examples (the author's words in the thread: "Just let the AI pip install this library, since it comes with lots of examples").
This differentiation is one reason it took off. The two bigger factors, though, are that the examples target the shields domestic users struggle with every day, and that the author brings an existing "Ruyi" brand. Technical positioning explains why people star and fork; the traffic itself came from the existing training audience and linux.do.
5. Visibility and copycats
- Visibility: it's "circulating in small circles." The Chinese reverse-engineering/scraping community (linux.do, Bilibili, Official Account, WeChat groups) is spreading it. English X, HN and Reddit are basically silent. No media coverage. It doesn't charge, so there's no public revenue.
- Copycats: none spotted directly yet. Sibling repos under the same brand already exist: language ports of ruyiPage (ruyipage-go, ruyipage-js, etc., with a dozen or so stars each) and Facetomyself/ruyi-mcp. The name-alike icesword0760/ruyi (remote browser + AI workbench) is unrelated.
6. What others can borrow
- Make a "working bypass example for X shield" the centerpiece of the README, so users can see at a glance whether their shield is covered.
- Launch first on a technical community like linux.do using the "open source promotion" format, reply to every comment in the thread, and add cases live.
- Use the open source tool to attract a sponsor (an API relay), funnel people into WeChat groups, and monetize through two channels: AGPL plus a commercial license.
- When a repo gets taken down, immediately re-create it under a new account and remove the sensitive targets. The original users will track it down via the post, and many will fork copies to keep.
VII. What I'm still unsure about
- Why the original repo was 404 (a lawyer's letter, a platform report, or the author deleting it themselves), who owned it, and how many stars it had, none of that could be confirmed.
- The GitHub stargazers endpoint (REST and GraphQL) returns an empty list or 404 for this repo, so the quality of the star accounts can't be spot-checked. The cause is unknown; GitHub may be restricting this repo. The fork accounts sampled look like real people.
- Not verified whether the domestic-target examples that were removed were later shared privately in WeChat groups or courses.
- Not clear how many PyPI downloads come from mirrors or CI.
Sources
- GitHub API: repos/cutetailbud/ff-ruyijs (repo, commits, releases, issues, forks), 2026-11-06
- linux.do launch post and replies: linux.do/t/topic/3078486
- Trendshift:trendshift.io/repositories/325142
- pypistats:pypistats.org/api/packages/ruyijs/recent
- Xiaojianbang Education: 0xshoulderlab.site/; Esword618/ruyiBrowser
- Tencent News: news.qq.com/rain/a/20261029A005GR00; view.inews.qq.com/a/20260623A09GUW00
- Bilibili: BV13rno6PEvn, BV1mmxY6zEV7 (play counts from the search results page)
- Data card: /home/ljr/fanout/why/cards/gh_cutetailbud__ff-ruyijs.json