SEO/AEO: one structured-data graph, llms.txt, AI crawlers named

[Abuubkar]

The search half of SEO was done in #38. This is the answer-engine half — what
ChatGPT, Perplexity, Gemini and AI Overviews read, which is not the same set of
things Googlebot reads.

Structured data is now one graph

It was two islands: Organization from the root on every page, FAQPage from
the home route, neither carrying an @id, so no node referred to any other.

Now the root emits Organization + WebSite with stable @ids, and each route
adds its own nodes pointing at them:

Route	Adds
/	WebPage, Service (with the Services section as a nested OfferCatalog), FAQPage
/privacy, /terms, /hipaa	WebPage with dateModified, BreadcrumbList

Consumers merge a graph by @id across separate <script> tags, so the
publisher is described once and everything else refers to it.

New fields, all derived from messages.json — nothing invented:

• Organization.knowsAbout — the specialty list. This is the field an engine
  asked "who bills for dermatology practices" has to match on.
• Organization.contactPoint, email
• Service.hasOfferCatalog — the four Services cards, nested by stage of the
  revenue cycle, each line item as an Offer
• Service.audience
• WebPage.dateModified on the legal pages, parsed from the one date already in
  legal.json rather than stored a second time

Two keys added to messages.json: meta.serviceType and meta.audience.

public/llms.txt

The landing page restated as plain text an answer engine can quote: the
positioning bullets, services with every line item, the four steps, specialties,
supported EHRs, the full Why Us comparison table, what gets reported monthly,
the fit/not-fit lists, all eight FAQ answers, and the four pages.

Generated by scripts/llms-txt.mjs, run from pnpm build. Generated rather
than written because a second hand-kept copy of these facts is exactly the drift
that put a wrong logo in the WhatsApp preview. The output is committed, so a
copy change shows up in the diff.

robots.txt names the AI crawlers

This is the finding worth reading. https://execil.net/robots.txt does not
serve the file in public/ — Cloudflare prepends a managed block that
disallows GPTBot, ClaudeBot, Google-Extended, Applebot-Extended,
meta-externalagent, Amazonbot, CCBot and Bytespider, and sets ai-train=no.
Default-on, not a decision anyone made here.

Everything else in this PR is decoration while that stands. Filed as #43 with
the dashboard steps — it needs the Cloudflare account, so no agent can do it.

public/robots.txt now names each of those crawlers in an explicit Allow
group. Repeating a user-agent merges the groups and an equal-specificity Allow
wins the tie-break, so for crawlers that follow it the block is already
reversed. A fallback, not the fix.

FAQ anchors

Each answer gets id="faq-<slug>" and a matching @id on its Question node,
so one answer can be linked and cited on its own.

Verified

• pnpm check clean, pnpm build prerenders all four pages
• Home emits 5 nodes across 2 blocks, legal pages 4 — @ids resolve, no
  duplicated Organization
• dateModified: "2026-09-08" on /privacy, matching the page's own date
• Live crawler access re-checked: GPTBot, ClaudeBot, PerplexityBot,
  OAI-SearchBot, Googlebot and bingbot all get 200 and the full page — the
  block is in robots.txt only, nothing at the edge
• https://www.execil.net/ still answers 200, not 301 (SEO: the three items no agent can do #39 item 2, unchanged)

[cloudflare-workers-and-pages]

Deploying with    Cloudflare Workers

The latest updates on your project. Learn more about integrating Git with Workers.

Status	Name	Latest Commit	Preview URL	Updated (UTC)
✅ Deployment successful! View logs	execil	8239be7	Commit Preview URL Branch Preview URL	Sep 13 2026, 05:05 AM