# AI Website Pipeline

> One keyword in, one small static utility site out. A fully automated agentic pipeline that runs on
> your machine, not ours: 13 layer contracts, 395 machine-evaluable gates, a working CLI and a
> domain engine. The method, the niche score calculator and both published datasets are free.
> The repository is 199 USD one time, and 499 USD adds a niche screen and a written review of
> your first built site.

The site's argument is that nothing between the keyword and the deploy can tell you no, and that the
missing part is a machine that says no. It publishes Search Console data for four sites the author
runs, invites you to compare the failure column first, and states that the demo which fails is the
proof, a gate that refuses to assume twenty absent flags are clean. Subjects it documents include
RDAP, keyword difficulty distribution, keyword maps, information gain, Search Console coverage,
IndexNow, GEO sprint stages and JSON Schema.

## Pages

- [Home](https://aiwebsitepipeline.com/): the argument, the four sites and their Search Console data,
  the failing demo, what 199 USD buys, and who built it.
- [Pricing](https://aiwebsitepipeline.com/pricing/): three tiers. Read the spec at 0 USD, the
  pipeline at 199 USD one time, and 499 USD for the pipeline plus a niche screen and a written
  review of your first built site within three business days. Sixty day refund on both paid tiers.
  At 499 USD the download link is emailed within one business day rather than released instantly.
- [Method](https://aiwebsitepipeline.com/method.html): how the pipeline runs.
- [Niche score](https://aiwebsitepipeline.com/niche-score.html): how a niche is scored before build.
- [The .com namespace is full](https://aiwebsitepipeline.com/the-com-namespace-is-full.html): essay
  on domain availability.
- [Click yield](https://aiwebsitepipeline.com/page-yield.html): a publisher side Search
  Console census of 35 small static sites. Of 5,177 pages Google showed over 90 days, 9.04
  percent earned at least one click, the median site sits at 9.57 percent, and the per
  domain spread runs from 0 to 50 percent. Raw JSON published beside it, plus a browser
  tool that places your own two Search Console counts in the measured distribution.
- [FAQ](https://aiwebsitepipeline.com/faq.html)
- [Order lookup](https://aiwebsitepipeline.com/order-lookup.html)
- [Twitter card validator](https://aiwebsitepipeline.com/twitter-card-validator.html): paste HTML or fetch an https URL, see the card X resolves with the Open Graph fallback applied, the real image measured, and every rule graded against the line it comes from. Census of 58 major homepages beside it.
- [Open Graph checker](https://aiwebsitepipeline.com/open-graph-checker.html): paste HTML or fetch an https URL, check the four required og properties, structured image properties, type vocabulary and locale format, with the image loaded and measured. Census of 58 major homepages beside it.
- [Hreflang checker](https://aiwebsitepipeline.com/hreflang-checker.html): paste a page, a sitemap fragment or plain code and URL lines, or fetch an https URL, validate every code against the IANA registry, check x-default and self reference, and fetch the alternates to confirm return links. Census of 58 major homepages beside it.
- [Keyword density checker](https://aiwebsitepipeline.com/keyword-density-checker.html): paste text or HTML, or fetch an https URL, count words, top terms and phrases with a published stop word list, and see the density real homepages have, measured on 48 major sites, with no target recommended.
- [Terms](https://aiwebsitepipeline.com/terms.html)

## Open-source libraries published alongside it

- [gate-contract](https://docs.rs/gate-contract/): Rust library for fail-closed gate contracts. A
  blocking gate halts the run both when it fails and when it cannot be evaluated, and a Known<T> type
  propagates Unknown instead of decaying to a default.
- [indexnowkit](https://code.dlang.org/packages/indexnowkit): D library that builds and validates
  IndexNow submissions against the published protocol constraints.

## Notes for retrieval

- Everything on this domain is open to crawlers except /dl/, which holds delivered product archives
  behind unguessable paths.
- Author: Michael Lip (https://github.com/theluckystrike).