Project Everywhere Green Campaign: Ten Brands, Thirteen Sites, One Friend's Domains
A Friend Handed Over a Batch of Domains. Ten Brands Got Built Across Them in Four Days.
The audit found more than it expected to.
What was measured.
Three readings from four days. The launch count is the headline number and the least interesting one.
Fast enough to be the headline. Speed is not what made it work, and on its own it is the number most likely to hide a problem.
Day one caught an incomplete build before it multiplied across the batch. Day four caught two more bugs that day one's own fix had produced. Every stage checked its own work.
Verified entity count is what separates a built network from a launched one. Six sites with nothing behind them would have counted the same in the launch total.
One property in the batch, TheGreenThread.org, was confirmed unbuilt rather than quietly counted as live.
The short version.
Becky, the lawyer friend behind FloridaOnAppeal.com, contributed a batch of domains for a Green Living site and a handful of small-business editorial and directory ideas. What came back was thirteen live properties across six brands: a two-site Green Living pair, a two-site health-discovery pair, and a four-brand small-business run that included a six-site Atlanta directory network, all shipped between August 17th and August 20th, 2026, plus a next-day audit that caught real bugs in two of the freshest builds.
Explain it to me like I'm ten.
A friend gave us a bunch of website ideas. In four days we built 13 real, live websites: two about green living, two about health, and several small-business sites including a whole Atlanta food-and-nightlife directory. The next day we double-checked our own work and found (and fixed) two real bugs in the newest ones.
What was the system?
The first build, KeenlyGreen.com, didn't start clean. An earlier Codex pass had shipped it without any of the portfolio's DK Federation v8.0 scaffolding: an empty config layer, three redundant AI-discovery files where the spec wants one, and /brands/ and /materials/ hub pages silently rendering the wrong content because no brand or material data existed yet. The campaign's real starting point was an audit, not a launch.
What changed?
KeenlyGreen.com got rebuilt from the scaffolding up on August 17th: a bootstrap.php routing engine, a real config layer, structured investigation and brand/material content, a working contact form replacing a hardcoded plaintext email, and full DK v8.0 federation output. It became the founding site of a new Green Living constellation. KeenlyGreen.org followed the same day as its evidence-layer sibling, completing a two-site pair.
August 18th shipped five major builds in one day: HealthIsCalling.com and HealthIsCalling.org as a consumer-discovery and evidence-library pair, BellyUp ATL Phase 1 (bellyupatl.com plus vegan, BBQ, and bars audience-lens subdomains sharing one venue entity store), HoldMyBagel.com, TheModernWriter.com, and WomanVersus.com, each an independent editorial platform built from scratch. BellyUp ATL's Phase 2 followed on the 19th, adding family and nightlife subdomains on the same shared codebase, bringing that one network alone to six live sites.
Each build kept the portfolio's honesty discipline intact under real time pressure. BellyUp ATL's launch deliberately shipped honest empty states on its UGC feed and community-quotes sections rather than fabricate testimonials to make day one look populated. Research caught and excluded a venue from the original brief that had actually closed, and corrected several wrong neighborhood assignments before any venue JSON got written. Phase 2's new experience-attribute filters only ever surface options at least one real venue is actually tagged with, rather than offering filters the data can't back up.
August 20th closed the arc with an audit, not a build: a portfolio-wide check for a hardcoded-sitemap-list bug already fixed elsewhere found two real instances in the fresh batch.
What did the evidence show?
The audit found more than it expected to. HoldMyBagel.com had the straightforward version of the bug: real pages hardcoded as individual add() calls instead of loaded from content. BellyUpAtl.com's base site had a subtler version: its page schema has no status or include_in_sitemap field at all (every record is always live), so the fix already used everywhere else in the portfolio (filter on that field) would have silently dropped both of that site's existing pages instead of fixing anything. Reading the site's own routing code before applying the fix, rather than assuming every property shares one schema, was the difference between a real fix and a new outage.
The audit also settled an open question: TheGreenThread.org, referenced in the Green Living constellation's own planning docs, has no local project folder and no live docroot anywhere on the VPS. It is not built. Naming a property in a plan doesn't make it exist, and the audit said so instead of assuming the plan was already reality.
The scale is the other finding. Five site launches logged in a single calendar day, August 18th, across four unrelated brand concepts, is not a pace the campaign set out to prove. It's just what the actual DK Warehouse event log shows happened once the KeenlyGreen scaffolding problem was solved on day one and the same architecture pattern could repeat.
What can be reused?
Thirteen live sites in four days sounds like the story. The actual story is that day one was an audit that caught an incomplete build before it multiplied, and day four was another audit that caught two more bugs in the batch day one's fix produced. The campaign moved fast because every stage checked its own work.
Keep Browsing the Case Library
Move from this evidence file back into the full proof system.