Stringers
Pluviobyte/saas-growth-agents · Pluviobyte

13-day 525-star "SaaS growth agent"; the four hundred stars came over from a rename

What it is It packages the daily growth chores of an overseas website (finding new keywords, checking competitor backlinks, submitting to directory sites, archiving good social posts) into SOPs and scripts that agents like Grok Bot or Claude Code can install and run on a schedule. It also includes a small tool for scraping Douyin comment sections.

Why it took off The author's X account, with 53k followers, tweeted within two minutes of creating the repo. "Agent scrapes Douyin comments without a captcha" hit a real pain point and brought in about 400 stars from 159k views. Later, an old repo was renamed to ride the Grok Bot wave, and the stars carried over as-is. A long-form article then brought in about 100 more.

CHow you could follow Stop making yet another growth-SOP collection. Copy how it shipped: pick one concrete pain point that agent users hit every day, build a small, precise skill with verifiable test results and a one-line install, and launch the tweet and the repo together. When the buzz fades, rename the old repo and fold it into the next hot topic.

C grade: both star waves came from the author's own large account, which followers can't borrow. The really differentiated Douyin comment tool wave has already been harvested. In the growth-SOP space there's an English collection with 60k stars and a deeper Chinese new-keyword radar, and it's crowded. You can learn the playbook, but following it isn't worth it. Whispers in the niche
Data (updated daily) · Repo created 2026-10-30 · 551 stars · 7 days +20
GitHub · stars
40045050055011-1011-1211-1511-1811-20551
Data 2026-11-10 → 2026-11-20 (11 days, 11 observations). Points before 10-28 come from deeply / archive backfill, not every day has one

saas-growth-agents is an open-source repo from a Chinese AI blogger. It packages the daily growth chores of an overseas website (finding new keywords, checking competitor backlinks, submitting to directory sites) into SOPs that agents like Grok Bot or Claude Code can install and run on a schedule. On the surface, it got 525 stars and 128 forks in 13 days. In reality, roughly 80% of the stars were earned by its predecessor. The predecessor was a small tool for scraping Douyin comment sections. The author, Nov 10, renamed the repo, and the stars came along unchanged. Below: where the two waves of stars came from, which move you can learn, and which one you can't borrow.

One tweet, two minutes

2026 year Oct 30 15:47 (UTC), a new GitHub repo called douyin-comments appeared. Two minutes later, X account @Pluvio9yte posted a tweet:

The problem of agents constantly getting captcha'd when scraping Douyin comment sections has finally been fully solved by a skill. This skill scrapes every comment in a Douyin comment section without triggering captchas, about 2.5 times faster than manually clicking to expand replies with browser automation. What you can do: 1. Scrape pain points, find demand, see what people complain about, then find business opportunities 2. If you're making a video, check what people ask in the comments of competitor videos...

The tweet got 159k views, 942 likes and 1,785 bookmarks (11-11 on X). One replier with 55k followers wrote: "Holy crap, this is awesome, I often read comment sections to find demand."

Building the repo and posting the tweet were only two minutes apart, which means the tweet was written in advance and posted as soon as the repo existed. The author has 53k followers and positions as "AI, vibe coding, agents, going global with websites". In the next 48 hours the repo picked up 87 forks. By Nov 8, forks had reached 105 and then mostly stopped. GitHub does not give day-by-day star counts for that period, but the snapshot on the day of the Nov 10 rename was 422 stars and 105 forks. From that, it's inferred that roughly 400 stars had already accumulated before the rename, all from this tweet.

The tool itself did have substance. Instead of cracking Douyin's request signature, it lets the agent replay the real requests the page sends inside your own logged-in Chrome. The signature is computed by Douyin's own page code, so no captcha appears, and nested replies can all be scraped. The author also posted a manual check of the results: 195 correct out of 195 items.

The rename: moving the old repo's stars onto a new hot topic

On the morning of Nov 10, the author did two things.

First: one commit, "Become saas-growth-agents". The repo gets renamed, a full overseas-growth SOP goes in, and the Douyin comment tool moves into a subfolder. GitHub redirects the old URL 301 to the new one, so stars and forks come along. The README has a line at the top: "Coming from douyin-comments? This repo *is* the old douyin-comments."

Second: published a long-form article on X titled "Build Your Personal Agent into an Automated Money-Making System". The opening reads:

Many people's first question about Grok Bot is the same: can it help me make money? My answer: no. ...How to turn your repetitive work into SOPs, then hand them to a Bot to run.

The timing was deliberate. xAI launched Grok Bot in Aug 11, with daily active users passing 600k by late October (Phoenix Tech). Meta's Muse in September and OpenAI's dots in late September also launched around then. In that period, the most common question from Chinese-speaking users with an always-on personal agent was whether they could use it to make money. The long article got 93.5k views and 1,417 bookmarks (11-11 on X). The same day, an English account with 9.2K followers posted a list, "The best Grok Bot setups are hidden in these 20 GitHub repos", and ranked it 4th.

Here is how it played out:

DateStars / forksWhat happened
10-300Created the repo as douyin-comments, tweeted two minutes later
10-30 to 10-31+87 forkFirst wave, est. 400 stars
11-01 to 11-08+18 forkNatural decline
11-10422 / 105Renamed, published the long article
11-11494 / 121+72 stars, the biggest single day after the rename
11-12525 / 128+31

Data from GitHub. The two posts' view counts were close (159k vs 93k), but the stars they brought differed by about 4 times (about 400 vs about 100). Two reasons, both inferred. First, the long article was itself substantive, so readers just bookmarked it and didn't need to click into the repo. Second, the barriers differed: the Douyin tool could be cloned and used immediately, while the growth system required first having an agent such as Grok Bot and then getting a paid data API key.

One more thing: of the 128 forks, 126 are empty forks with zero changes, more like bookmarks or backups. Fork count alone doesn't show anyone is building on top of it.

Where the money comes back

The author doesn't mention revenue, but the README is blunt about it. The recommended agent is Grok Bot, with the author's invite link next to it: users get 14 days free, and the README says in plain words "This is a referral link, and the author is rewarded for sign-ups." All external data goes through AIsa, a pay-per-call API aggregator, and that link is the author's personal one too. Step three of the install is "go get an AIsa key and put it in the Bot's vault," so everyone who follows the setup passes through both links.

AIsa is systematically handing out personal links to skill authors and bloggers. Another repo, a "new-term radar", also carries its links. How much they actually pay back, and whether the author gets sponsorship money on top, has not been disclosed.

Who it's competing with

All three categories have projects an order of magnitude bigger:

ProjectLaunchedScale (GitHub, 11-12)How it differs
NanmiCoder/MediaCrawler2023-0670,884★Multi-platform crawler, needs deployment; this is an agent skill, zero deployment, and it doesn't trigger captchas
coreyhaines31/marketingskills2026-0159,743★Full English marketing skill set, ask once and get one answer; this is a pipeline that runs on a schedule every day
AgriciDaniel/claude-seo2026-0228,171★Leans toward SEO audits, does not do new-term discovery or directory submissions
zhuyansen/new-term-radar2026-10-122,163★ / 644 forkSame "new-term radar" idea, with more signals and backtesting on each one; also carries AIsa links
flaqai/backlink_skills2026-08888★Only submits to directory sites for backlinks
unicodef1wn/grokbot-field-notes and others2026-08 to 10490–701★General Grok Bot configs and checklists, not focused on going-global growth

On growth it has no obvious product edge. Its pitch is "the SOP the author uses for their own going-global sites" and "say one sentence, and the Bot installs itself." The real differentiator is the Douyin comments tool, and that wave of stars was already spent.

The author is a reliable star-printing machine

Look at the author's other repos to see the pattern: rnskill 1,737 stars, video-production-skills 705 stars, Kiro-auto-register 321 stars, claude-ban-clean 288 stars (10-29 created, 10-31 gained +80 the same day). 8 repos together come to roughly 4,000 stars (GitHub). The playbook is the same every time: every week or two, ship a Chinese skill that solves one concrete problem, push it with one tweet that names the pain point, and collect 100 to 500 stars per push.

saas-growth-agents is another output of this machine, plus one new move: old repos whose heat has faded don't get thrown away. They get renamed and attached to the next hot topic.

If it were you

Verdict: you can learn a launch playbook here, but you won't get these stars. Both star waves came from the author's X account with 53k followers, and the Douyin comments tool, the one with real differentiation, has already had its wave. The growth-SOP space is crowded: there's a 60k-star English compilation, and a deeper Chinese new-term radar. Copying this into another "going-global growth agent compilation" without your own traffic will most likely land in the dozens of stars.

Who it's for: developers with a tech account of a few thousand to tens of thousands of followers, or anyone serious about building an account, who wants to learn how to launch a small tool and pull in hundreds of stars at once. Also people sitting on an older repo whose heat is fading but still has a few hundred stars, who want to attach it to a new hot topic.

Who it's not for: anyone with no traffic of their own who thinks "packaging growth chores as agent SOPs" is enough to earn stars by itself. People who want to scrape Douyin or Xiaohongshu should also think twice: these tools live on the edge of what platforms tolerate, and any interface change from the platform means rewriting them.

How to do it: copy the structure, not the content. Pick one concrete snag that agent users hit every day (here, "scraping Douyin comments keeps triggering captchas"), and build a small, precise skill, not a sprawling one. Include verifiable real-world test results (like the 195/195 shown here). Give the README a one-sentence install path (bootstrap.md plus AGENTS.md) so people who don't write code can use it. Once the heat fades, fold it into a larger repo that rides the next hot topic, and put "coming from XX? start here" at the top of the README. If you attach referral links, do what they did and disclose them clearly.

How to get users: write the tweet first, then create the repo, and publish both within two minutes. The first sentence of the tweet names the pain point, the second gives the result and a number, the third says what you can do with it, and the link goes last. Post Chinese versions to X and WeChat Official Accounts; post English versions to X and the matching subreddits. When a new agent platform takes off, a wave of English list-style posts shows up (here, the Grok Bot config lists). Proactively send them your repo.

When to move: this personal-agent wave (Grok Bot, Muse, dots) started in August and is hot now. There are already plenty of config and template repos, but most have only a few hundred stars. To ride it, you'd want a skill that solves one concrete snag out by 2027 January. If you're still building "general config compilations" by then, you're late.

Signs you're reading this wrong: ① a going-global growth skill repo with no big-name backing passes 1,000 stars in a month, which would mean the category can earn stars on its own, not just through the author's traffic; ② during weeks when the author posts nothing new, saas-growth-agents still holds above 100 new stars per week, which would mean people are actively searching for and using the growth system itself; ③ Grok Bot launches an official skill marketplace that gives exposure based on install counts, so distribution stops depending only on bloggers and people without followers get a shot.

GitHub repo ↗

Research notes: competitor table, acquisition evidence, full research report
Competitors
NameWhereLaunchedScaleHow it differs from this one
NanmiCoder/MediaCrawlerOpen source2023-0670,884★Crawls multiple platforms with Playwright, needs deployment. This repo is an agent skill: it copies real requests from an already logged-in Chrome, signs them with the page's own SDK, and doesn't trigger captchas.
Evil0ctal/Douyin_TikTok_Download_APIOpen source2021-1121,354★Self-hosted API and MCP, mostly for video downloads, needs Docker deployment. This repo only scrapes comment sections, zero deployment.
TikHub / Apify Douyin comments scraper / n8n + JustOneAPIOther—Paid APIPay-per-call. The free tier of Apify gives only 50 comments. This repo is free and can pull every nested reply.
coreyhaines31/marketingskillsOpen source2026-01-1559,743★Full English marketing skill set, one-off consulting-style. This repo is a growth pipeline that runs on a schedule every day, with a ledger and approvals.
AgriciDaniel/claude-seoOpen source2026-02-0728,171★Leans toward SEO audits. This repo does new-term discovery, competitor backlinks, and directory submissions.
Ryze-AI-Adgent/open-seo-mcp-skillsOpen source2026-08-299,288★A set of SEO MCP tools. This repo is an SOP for an always-on Bot.
zhuyansen/new-term-radarOpen source2026-10-122,163★ / 644 forkSame concept, and the new-term radar is more developed there (a dozen-plus signals plus backtesting). Here the new-term radar is only one of three Bots, alongside site operations and a social media kit.
flaqai/backlink_skillsOpen source2026-08-17888★Only submits to free directory sites. This repo folds that into one step of a site-operations Bot.
unicodef1wn/grokbot-field-notes and other Grok Bot config reposOpen source2026-08 to 10490 to 701★ (field-notes 701, hypergrok-trading-desk 492, awesome-grokbot-playbooks 491)General role, trading, or resource lists. This repo focuses on going-global site growth, with Chinese-language SOPs.
The AI features built into Ahrefs / SemrushBig-company features—SaaS subscriptionsThey require a paid subscription, and you can't hand them to your own agent to run every day automatically.
Where users come from
  • Verified The author's own X tweet (first posted with douyin-comments): @Pluvio9yte, 53k followers, posted at 10-30 15:49 UTC, only 2 minutes after the repo was created: 159k views, 942 likes, 1,785 bookmarks. 87 forks clustered between 10-30 and 10-31.
  • Verified Old repo renamed, inherits stars: 11-10 05:05 committed "Become saas-growth-agents". The old address douyin-comments 301 redirects here, and the README says 'This repo is the old douyin-comments'. On the day of the rename, the snapshot already showed 422 stars / 105 forks.
  • Verified The author's X long-form post (riding the Grok Bot and other personal-agent wave): 11-10 X Article "Turn Your Personal Agent into an Automated Money-Making System": 93.5k views, 1,417 bookmarks; the quote tweet got 14.5k views. Afterward, 11-11 gained +72 stars and +23 forks in a single day.
  • Verified An English list-post in the Grok Bot ecosystem: @unicodef1wn posted "11-10 best Grok Bot config repos"; this repo ranked #20, with 4,4.2K views and 104 bookmarks.
  • Verified Passive indexing by skill aggregators and directories: SkillsLLM, AgentSkillsHub (Best Claude SEO Skills), github.day, RepoRadar, and TypingMind all have pages for it.
  • Verified Referral funnel (monetization, not acquisition): the README states that the Grok Bot invite link (14 days free) and the AIsa link go.aisa.one/rain are the author's referral links. AIsa issues creators personalized links (new-term-radar uses go.aisa.one/yansen-zhu).
  • Inferred Recommended by @wohsj110: the recall card credits the author's tweet, but it was actually another person (Aiden). The two tweets got only 59 and 34 views, with zero likes, so their effect is negligible.

Why Pluviobyte/saas-growth-agents took off

Bottom line: this repo's stars were not earned by "SaaS growth agents." Roughly 80% came from its predecessor, douyin-comments. Two minutes after creating the repo, the author "Pluvio" (@Pluvio9yte, 53k X followers) posted a tweet that got 159k views and 1,785 bookmarks, bringing 87 forks and about 400 stars within two days. 11-10 The author then renamed the repo, added three going-global site-operations SOPs, and published an X long-form post, "Turn Your Personal Agent into an Automated Money-Making System" (93k views). That move rode the Grok Bot / Muse / OpenAI dots wave of always-on personal agents, and the README carries both a Grok Bot invite link and an AIsa referral link. The second wave added only about 100 stars. Essentially, this is a Chinese AI blogger's playbook: take a repo that already has stars, rewrite it as a traffic entry point for a new hot topic, and add a referral funnel.

---

1. What it is

  • Repo: github.com/Pluviobyte/saas-growth-agents, MIT, Python (standard library only), 125 files. Created 2026-10-30 15:47 UTC.
  • Current positioning: the daily growth chores of going-global SaaS, written as SOPs and scripts an agent can execute, split into three Bots:
    • New-term radar: three reports a day, with data from Google Trends, competitor sitemaps, new listings on directory sites, Product Hunt, GitHub Trending, Reddit, and TikTok.
    • Site operations: finds competing sites by keyword, checks their DR and backlinks, and submits to free directory sites.
    • Social media analysis kit: hand it a link, and it fetches the original, summarizes it, and files it into a knowledge base. The original Douyin comments scraping skill lives at bots/social-kb/douyin-comments/.
    • There is also a trading radar for analysis only, with no order execution.
  • The install is "send one sentence to your Bot": ask Grok Bot, Muse, Dot, Hermes, OpenClaw, Claude Code, or Codex to read AGENTS.md and install itself (bootstrap.md).
  • All external data goes through AIsa, a pay-per-call API aggregator. The README states that the AIsa and Grok Bot links are both the author's referral links: go.aisa.one/rain and grok.com/refer/bot?code=… (README original: "This is a referral link and the author is rewarded for sign-ups").
  • The commit history has only 5 entries (GitHub API):
    • 10-30 15:47 Initial release: douyin-comments skill
    • 10-31 Added the verified demo video
    • 11-10 05:05 "Become saas-growth-agents" (renamed, folded into the growth system)
    • 11-10 Two small edits followed, unifying the AIsa signup link to go.aisa.one/rain
  • The old address github.com/Pluviobyte/douyin-comments 301 redirects here, and both stars and forks carried over.

2. Growth data

TimelineEventStars / forksSource
10-30 15:47 UTCCreated the repo, originally named douyin-comments0GitHub API
10-30 15:49 UTCThe author posted on X: "The problem of agents scraping Douyin comments and constantly triggering captchas has finally been completely solved by a skill"—x.com/Pluvio9yte/status/2116195906328637534:159k views, 942 likes, 111 reposts, 1,785 bookmarks, 96 replies
10-30 to 10-31First wavefork 32 + 55 = 87GitHub forks API (counted by fork creation time)
11-01 to 11-08Natural decayonly 18 more forksSame as above
11-10 (first time our database caught it)the day of the rename422 stars / 105 forksgithub repo_day
11-10 01:57 UTCauthor posted a long X article, "Turn Your Personal Agent into an Automated Money-Making System"—x.com/Pluvio9yte/status/2119972735426383993:93.5k views, 723 likes, 1,417 bookmarks; same-day quote post with 14.5k views
11-10 12:27 UTC@unicodef1wn (9.2K followers) posted a list titled "The best Grok Bot configs are hidden in these 20 GitHub repos." This repo ranked #4 (469★ at the time)—x.com/unicodef1wn/status/2120131268092297648:4.2 — 104K views, bookmarks
11-11Second wave494 stars (+72) / 121 forksGitHub repo
11-12525 stars (+31) / 128 forksGitHub repo

Notes:

  • Our github database didn't record this repo under the name douyin-comments before the rename, so there are no day-by-day star counts between 10-30 and 11-09. The GitHub stargazers endpoint returned 404 this time and doesn't expose star timestamps. The forks endpoint does show when each fork was created: 11-08 forks total as of 105, matching the 11-10 snapshot's 105 forks. From that we infer roughly 400 stars existed before the rename, coming from the Douyin comments skill wave.
  • The two tweets got similar views (159k vs 93k), but the second drove about 4x the stars (roughly 400 vs 100). The long article is the content itself, so readers bookmarked the article (1,417 bookmarks) instead of clicking through to the repo. Using the system also requires an agent like Grok Bot plus a paid AIsa key, a much higher bar than "git clone a Douyin comments skill."
  • The recall card says "the author 11-01 promoted it on X (@wohsj110)", which is wrong. @wohsj110 (Aiden, 2.2K followers, Xiamen) is a different person. The two English posts he made got just 59 and 34 views, with zero likes, respectively, 11-01. The one that actually worked was the author's own account, @Pluvio9yte.

3. Where users come from (growth tactics)

  1. The author's own X account is the main channel. @Pluvio9yte ("雪踏乌云") has 53,139 followers and positions himself as "AI & Vibe coding & Prompt & Agent, going global via the web." The long article was also published on WeChat Official Account. Both traffic waves came from him: he tweeted two minutes after creating the repo, which means the tweet was written first and the repo was created second, a planned launch.
  2. A repeatable playbook: a Chinese skill repo plus a viral tweet, with an audience already in place. Repos under his name:
    • rnskill 1,737★(2026-07),video-production-skills 705★,Kiro-auto-register 321★
    • claude-ban-clean 288★ (created 10-29, released the day before, +10-31★ on 80)
    • dot2api 216★,awesome-talking-head-skills 103★
    • These numbers come from the GitHub API and our github database. Every week or two he releases a new Chinese skill repo and pushes it with an X tweet. This account is an engine that reliably produces 100 to 500 stars per launch.
  3. Rename to inherit stars and ride a trend. 11-10 renamed douyin-comments, a repo with about 420 stars and fading heat, to saas-growth-agents and folded new content into it. That way the new repo launches with 400 stars of social proof already in place, and it picks up traffic from the old link through the 301 redirect. The top of the README even says: "Coming from douyin-comments? This repo *is* the old douyin-comments".
  4. Ride the personal agent platform wave:
    • xAI launched Grok Bot 2026-08-11, and daily active users rose from under 100k in early September to over 10-26k 600 (Phoenix Tech); SpaceX's earnings disclosed 800k weekly paid users (Zhihu).
    • Meta launched Muse 9-08, and OpenAI launched dots 9-29.
    • The question new Chinese-language users ask most is "Can it make me money?" The author's long article answers it head-on, and the title is literally "Automated Money-Making System."
    • Included in the English list post from the Grok Bot ecosystem (unicodef1wn, #4). Lists like this bring a small amount of overseas traffic.
  5. Passive indexing by aggregator and directory sites: SkillsLLM, AgentSkillsHub (listed on the "Best Claude SEO Skills" page), github.day, RepoRadar, TypingMind skills, githubsignals, and others (see Jina search results). These are indexing events, not the cause of the spike.
  6. Where the money comes from (a recall hypothesis, which the README itself acknowledges): Grok Bot invite referral, where users get 14 free days and the author earns a reward, plus an AIsa per-call API creator link, go.aisa.one/rain. AIsa has a set of personalized creator links: another new-term radar repo, zhuyansen/new-term-radar, uses go.aisa.one/yansen-zhu, and a YouTuber uses go.aisa.one/ty-chen. That suggests AIsa is systematically recruiting skill authors and bloggers for distribution. Whether the author also received a separate AIsa sponsorship fee is unverified.

4. Competitors

Douyin comment scraping (source of the first wave of stars)

CompetitorsFormLaunchedScaleHow it differs
NanmiCoder/MediaCrawlerOpen source2023-0670,884★ (github repo 11-12)Uses Playwright to scrape multiple platforms and must be deployed as a full crawler. This repo is an agent skill: it copies the page's real requests directly from your logged-in Chrome and lets the page's SDK sign them itself (no a_bogus reverse engineering), so there are no captchas. Give Claude Code a link and it works.
Evil0ctal/Douyin_TikTok_Download_APIOpen source2021-1121,354★Self-hosted API and MCP, focused on video download and data collection, deployed with Docker
alphaply/DouyinComments, Yht20927/douyin-cli, and othersOpen source2024-02 / 2026-06211★ / 286★Scripts or CLIs, not agent skills
TikHub, Apify Douyin comment scraper, n8n + JustOneAPIPaid API——Pay per call; Apify's free tier gives only 50 comments. This repo is free and can also fetch every nested reply
Chrome extension "Douyin Comment Collector", and a comment-collection skill from Zhengdian AIChrome / other——Tools you click through, or hosted platform services that don't fit into an agent workflow

Agent skills for SaaS growth, SEO, and backlinks (the current stated positioning)

CompetitorsFormLaunchedScaleHow it differs
coreyhaines31/marketingskillsOpen source2026-01-1559,743★, +1,243 over 7 daysA full English marketing skill set (CRO, copywriting, SEO) that works as one-off consulting tooling. This repo is a scheduled pipeline (routine) that runs daily, with a ledger and approval gates
AgriciDaniel/claude-seoOpen source2026-02-0728,171★Leans toward SEO audits (26 sub-skills) and doesn't do new-term discovery or directory submission
Ryze-AI-Adgent/open-seo-mcp-skillsOpen source2026-08-299,288★A free SEO MCP plus skills, focused on keyword and ranking tools
zhuyansen/new-term-radarOpen source2026-10-122,163★ / 644 forkThe same "new-term radar" concept, but narrower and deeper (more than ten signal types, each backtested), and it also uses AIsa referral links. This repo's new-term radar is just one of three bots
flaqai/backlink_skillsOpen source2026-08-17888★Only submits backlinks to free directory sites
chenchaoge/seo-ai-linkOpen source2026-11-0724★An end-to-end backlink skill for Chinese-language independent sites, overlapping with site-ops
shipwords-dev/shipwordsOpen source2026-10-28366★A "new-term site network" build framework, downstream of the same going-global new-term pipeline
Yuzzyuk/marketing-os, marketing-skill-hub, thinkwithblack-skill, and othersOpen source2026-08 to 11200 to 575★Large marketing-skill collections
The AI features built into Ahrefs and SemrushBig-company features——Requires a paid SaaS subscription and can't be left for your own agent to run automatically every day

Grok Bot configs and template libraries (direct competitors in the second wave)

CompetitorsFormLaunchedScaleHow it differs
unicodef1wn/grokbot-field-notesOpen source2026-09-18701★Generic bot personas with failure logs, not focused on going-global growth
galleonlabs/hypergrok-trading-deskOpen source2026-08-16492★Trading-focused. This repo's trading radar is only a side feature
sit-in/awesome-grokbot-playbooksOpen source2026-10-17491★A Chinese-language list of Grok Bot plays
RongleCat/awesome-grok-bot, kydlikebtc/awesome-grokbotOpen source2026-08496★ / 482★Resource lists
gojiberryai-sales-os, grokbot-gtm-teamOpen source—196★ / 49★Sales and GTM

Bottom line: all three niches have plenty of competitors, and each has libraries an order of magnitude bigger. Its distinct edge in the "Douyin comments" niche is practical: it's an agent skill that copies real requests from your logged-in browser without triggering captchas, and in testing it got 195 of 195 comments right. In the "SaaS growth" niche it has no obvious product advantage. Its pitch is "the real SOP the author uses for his own going-global sites," plus "send one sentence and the bot installs itself."

5. Why it took off

  1. The author's account is the engine. A Chinese AI blogger with 53k followers who tweeted two minutes after creating the repo, with a top tweet at 159k views and 1,785 bookmarks. The fork timing shows 87 forks clustered within 48 hours of the tweet.
  2. The first wave hit a real pain point. People who create content or research demand have to browse Douyin comment sections, and agents keep hitting captchas when they try to scrape them. The tweet's opening line, "Finally fully solved by a skill," names that problem directly. A top reply said: "Holy cow, this is amazing, I often read comment sections to find demand" (@Saccc_c, 55k followers).
  3. The rename inherited the stars. The "new product" saas-growth-agents launched with about 420 stars and 105 forks, which made it look like "13 stars in 525 days." In reality, as a growth system, it earned only about 100 stars.
  4. Timing: Grok Bot, Muse and dots all launched within two months, and the Chinese dev scene is buzzing about whether personal agents can make money. The author caught that curiosity with a long post plus a "one sentence, bot installs itself" bootstrap (the long post got 93k views).
  5. Fork/star ratio of 0.24 is high, mostly because of the author's audience habits: rnskill has 0.12, Kiro-auto-register has 0.38. And of the 127 forks, 126 are empty forks of 0 stars (per the forks API), which looks more like bookmarking or backup than people actually building on it.

6. Differentiation, and whether it's the reason for the growth

  • What sets the Douyin comment skill apart: it doesn't reverse-engineer signatures. The agent reuses real requests inside the user's already-logged-in Chrome, so no captcha pops up, it pulls every nested reply, and it's about 2.5x faster than clicking through by hand. There was also a manual line-by-line check. This edge really was one of the reasons the first wave took off. The tweet focused on exactly the "no captcha" point, paired with the author's reach.
  • What sets the growth system apart: it's pitched as the author's real, in-house SOP. Three rules (scripts do deterministic work, agents handle judgment, a human makes the call); every bot shares one structure of job, routines and check files; it works with 6+ agents; and it bootstraps with one sentence. These are packaging and engineering-convention differences, not what drove the growth. The second wave of stars came mainly from the author's long post and the Grok Bot craze. The real competitor, zhuyansen/new-term-radar, goes deeper on the same concept and has more stars.

7. Visibility and copycats

  • Visibility: it's spreading in small circles. Mostly it moved on X among Chinese AI and going-global folks: the top tweet got 159k views, the long post 93k, plus one English Grok Bot list that reposted it once. No press coverage, no HN or Reddit, and it never trended (it was under 60 days old, so it couldn't have made the list). No public revenue, just the referral links the README discloses itself.
  • Copycats: I didn't see any repo directly copying it. Several libraries with the same concept appeared around the same time: zhuyansen/new-term-radar came earlier, and chenchaoge/seo-ai-link is a Chinese backlink skill from 11-07. The pattern of "Chinese growth skill library with an AIsa referral link" is already being run by at least two authors.

8. What to borrow

  • Tweet first, repo second, both published within two minutes: the opening line names a specific pain point ("my agent keeps hitting a captcha when scraping Douyin comments"), then gives the result (no captcha, 2.5x faster, open-source link).
  • Rename and inherit the stars: Retire a single-purpose tool whose buzz has faded, and fold it into a bigger system that matches the new trend. Old links will 301, and the README gets a line at the top: "Coming from XX? Start here." The new repo launches with social proof already in place.
  • Tying into referral programs of new platforms: new agent platforms (Grok Bot) have invite rewards, and data API aggregators (AIsa) have creator links. Writing both links into the install flow means that installing the thing already uses your referral link (the setup prompts users to register for an AIsa key). Disclosing the rewards in the README makes it look transparent rather than sneaky.
  • "One sentence and the bot installs itself": adding AGENTS.md to bootstrap.md makes it usable for non-developers too, which is a plus for new personal-agent users.

9. What's uncertain

  • I couldn't get daily star counts from before the rename: our database doesn't index the old name, and the GitHub stargazers API returned 404. The "about 400 stars before the rename" figure is inferred from the 11-10 snapshots and fork timing.
  • Whether the author cross-posted to WeChat Official Accounts, Xiaohongshu or Douyin, and how much that added, is unverified. His X bio says the long post first ran on his WeChat Official Account.
  • How much money the Grok Bot invites and AIsa referrals actually brought in, and whether AIsa paid any sponsorship fee, is unverified.
  • How many people actually run the bots after the switch to a growth system (real usage) is unknown. Stars and forks can't tell us that.

Sources

  • GitHub API: repo metadata, README, bootstrap.md, commit history, forks list (fetched 2026-11-11 to 12)
  • Our GitHub database (agent-im server, DuckDB): repo_day, repo_profile, data through 2026-11-12
  • X tweet data (api.fxtwitter.com): 2116195906328637534, 2119972735426383993, 2119973822548058307, 2116707233048457333, 2116706954790027395, 2116221894479466679, 2120131268092297648; author profile @Pluvio9yte
  • Grok Bot background: datacamp.com/zh/blog/grok-bot, tech.ifeng.com/c/8xfeSxEKOFE, zhuanlan.zhihu.com/p/2100877667004432893; Meta Muse: about.fb.com/news/2026/09/introducing-muse-personal-ai-agent/; OpenAI dots: openai.com/index/introducing-dots/
  • AIsa creator link: the README of zhuyansen/new-term-radar; youtube.com/watch?v=_XYvR_lLUmM

This article was written by AI from public data and the data we collect daily. Numbers are sourced, and inferences are labeled.