zsty.us

Before / After · Case Study

cbd.domains — the before/after engine forked into a high-risk showcase

  • Hemp / CBD

Cannabis case studies create an SEO problem for a general-purpose portfolio: high-risk content bleeds into how search and ad platforms classify the whole domain. The answer is segregation as architecture — cbd.domains is a full fork of the zsty.us engine (this site), rebranded by eight parallel agents working disjoint file clusters, holding the cannabis/CBD case-study catalog under its own legal entity. The cannabis /seo engine is live, with generated OG cards, sitemap/robots, Organization schema, hub cross-links, and llms.txt shipped. The showcase runs at cbd-domains.vercel.app until the apex domain is attached.

“The engine is the product twice over: once as this site, and once as the fork that keeps cannabis receipts from taxing everyone else's rankings.”
cbd-domains.vercel.app↗Repo → og-gitclaw/cbd-domains↗
🔒cbd-domains.vercel.app
jackiej.events — modern site

Receipts — measured

Numbers that moved.

0
Parallel rebrand agents (disjoint file clusters)
was 0
2
Engines maintained for two branded properties
was 2
1
High-risk case studies left on the general domain
was 1
  • Full engine fork: the zsty.us before/after machinery, rebranded end-to-end
  • 8 parallel rebrand agents over disjoint file clusters — no merge conflicts
  • Cannabis /seo engine live + OG cards, sitemap, org schema, llms.txt
  • SEO risk-segregation: high-risk case studies isolated from zsty.us

The story

Problem · Insight · Build · Outcome.

  1. 01 · Problem

    The best receipts were the riskiest to publish.

    The portfolio's deepest case studies — BMH, Buck Mountain, cbd.restaurant, Porto Verde — are cannabis. Publishing them all on the general-purpose domain meant the strongest work actively taxed every other property's search classification.

  2. 02 · Insight

    Fork the engine; segregate the risk; keep one codebase pattern.

    The before/after engine was already the portfolio's most valuable artifact. Forking it wholesale — rather than building a second site — means every engine improvement is portable, and the cannabis property starts at full feature parity on day one.

  3. 03 · Build

    Eight parallel agents, one rebrand, then the SEO layer.

    Fork of the zsty-us monorepo → 8-agent disjoint-cluster rebrand → cannabis /seo engine, generated OG cards, sitemap/robots, org schema, hub cross-links, llms.txt. Brand email on the self-hosted Stalwart server; the property sits under Candles and Clones LLC.

  4. 04 · Pending

    The apex is the last mile.

    cbd.domains (the apex) isn't yet attached to the Vercel project — it currently serves a stray lander — so the showcase runs at cbd-domains.vercel.app until the domain is added and DNS pointed. One console action away.

Architecture

One chain replaced by another. Receipts above.

Old stack
  • One domain, mixed risk
  • Cannabis + general content entangled
↓
New stack
  • zsty.us — general portfolio
  • cbd.domains — high-risk fork
  • Shared engine, separate entities
  • Cannabis /seo + OG + schema

What changed

Grouped by what kind of system shipped.

Each claim ships with concrete evidence — env vars, table names, cadence chips. No marketing fluff.

Agent backbone

An eight-agent parallel rebrand with zero merge conflicts

The fork was rebranded by eight agents running simultaneously, each owning a disjoint cluster of files — brand tokens, copy, metadata, SEO surfaces, assets — so the whole rebrand landed in one pass without a single conflicting edit. The engine fork itself is the case study: the same machinery that runs zsty.us, re-skinned in an afternoon of parallel work.

  • 8 agents, disjoint file-cluster ownership, one integration pass
  • Full brand swap: tokens, copy, OG, schema, llms.txt
Retention

Risk segregation as an SEO strategy

Search and ad platforms classify domains, not pages. Keeping cannabis case studies on their own domain under their own LLC means the general portfolio's rankings, ads eligibility, and email reputation never inherit the high-risk classification — and the cannabis showcase can lean into its niche without hedging.

Before
Every cannabis receipt published raised the whole domain's risk profile.
After
Two properties, two classifications, one shared engine.
Agent backbone

The cannabis /seo engine, live

The programmatic /seo machinery (sector × city pages) runs on the fork tuned for cannabis: sector pages live, generated OG cards per page, sitemap/robots, Organization schema, hub cross-links, and llms.txt for AI-crawler discovery — the full discovery surface, shipped in PR #1.

  • Cannabis /seo merged + live
  • Per-page generated OG cards; sitemap + robots + org schema + llms.txt

← All rebuilds

cbd.domains — the before/after engine forked into a high-risk showcase — zsty.us