Platform
The operating layer.
Nine stages, one account, one loop.
Research tools stop at telling you what is wrong. Builders stop at shipping. The operating layer restores, builds, ships, proves indexing and learns from the result, for one site or a portfolio of hundreds.
Stages and status
What exists today, and what is next.
We publish the status of every stage. Live means customers can use it. In build means it works internally and is being finished. Planned means it is designed and scheduled.
- 01
Decide
An idea or a niche in; sourced pain points, decisions and a risk-ordered build plan out.
In build - 02
Recover
Check an old domain and restore a capped vintage of it from public web archives.
Live - 03
Build
Site plan and pages from crawl-weighted patterns and a versioned design library.
Live - 04
Run
Host the site, ship ranked changes, check the live page, undo in one move.
Live - 05
Index
Prove each page is crawled and indexed, with verified search-engine visits and fix playbooks.
Live - 06
Distribute
Drafted social posts from site events, approved before they post.
Planned - 07
Measure
First-party, cookieless page analytics from edge logs.
In build - 08
Convert
Own forms capture leads; scored and answered within a minute.
In build - 09
Test
Page-split search tests and edge-split conversion tests, pooled across the fleet.
Planned
The change loop
Nothing ships unchecked. Nothing is one-way.
AI drafts and ranks changes. A person approves. The live page is checked after every release, and any release can be undone in one move.
The data loop
Each stage hands the next one something concrete.
| Hand-off | What passes |
|---|---|
| Decide → Build | A launch manifest: routes, sitemap, templates, schema types, required scripts, and the page plan. |
| Recover → Build | Restored pages run through the structure extractor and page-type classifier, so a dropped domain is rebuilt with the same URLs on the current library. |
| Build → Run | Versioned design and engine packages, plus a snapshot of what the system has learned about which structures work. |
| Run → Index | Every published page is watched: live URL, sitemap and Search Console binding. |
| Index, Measure, Convert → Build | Indexed, visited and converted, counted per page structure. This is how the system learns. |
| Test → Build and Decide | The winning variant updates the library; the assumption behind it is marked held or rejected. |
Architecture
Edge-native by design
Sites are served, changed and measured on a global edge network. Analytics are first-party and cookieless. Hosting cost per site is low enough to run large fleets profitably, and an operator can keep sites on their own edge account if they prefer.
Safety
The change loop
Read → choose → ship → check → undo. Every change is ranked by expected impact, approved, checked on the live page, and reversible in one move. Releases are versioned per site.
Proof
Verified crawling
Search-engine visits are verified with reverse and forward DNS, not trusted from user agents. Each site gets an indexation funnel and remedies for pages that stall.
Policy
What we will not build
No hiding that sites are related, no network link schemes, no promises of rankings. Uniqueness comes from content and structure gates, which is where search policy is heading anyway.
Two ways in
Start with an idea, or with a domain.
I have an idea or a niche
Decide turns it into sourced pain points, decisions and a risk-ordered build plan, then hands the plan to Build.
I have a domain
Recover checks restorability, archive age and topic fit, restores a capped vintage, and Build rebuilds it with the same URLs.
See the operating layer at work.
Talk to us about running your portfolio on the operating layer, or see how we prove it on our own sites.