Skip to main content
Q-Code Dev

SEO · AEO · LLMO · GEO

Built to be found

Every Q-Code site ships with the structure that search engines, answer engines, language models and local results each need, generated from the site rather than added afterwards.

A page has four audiences now and they want different things. The first wants something it can rank, the second something it can quote, the third something it can summarise, and the fourth to know where you are. None of this is an upgrade tier: it is in the template, so the first build has it.

Who reads a page

Four readers, four needs

SEO

Search engines

A page they can crawl, understand and rank.

  • A unique title and description on every page, within the lengths results actually show
  • One main heading per page and a heading outline that does not skip levels
  • A canonical link, so the same content on two addresses is not competing with itself
  • A generated sitemap with honest lastmod dates
AEO

Answer engines

A passage they can lift and attribute.

  • An answer capsule: the first paragraph after the heading answers the heading, in one self-contained sentence
  • Questions and answers marked up as FAQPage, generated from the entries you edit
  • Procedures marked up as HowTo, so steps can be read out in order
  • Facts stated in the HTML rather than assembled by a script after load
LLMO

Language models

A short, reliable summary of what this site is.

  • A generated llms.txt naming the site, its purpose and its key pages
  • Content signals in robots.txt declaring what may be used for search, input and training
  • One consistent voice across every written word, so the site reads as one author
  • No contradiction between what a page says and what its markup says
GEO

Local and generative results

To know where you are and whether to suggest you.

  • LocalBusiness structured data with address, hours and contact details
  • Review markup only where a real rating was given, never invented to fill the stars
  • An aggregate rating computed from every review, not the handful a page happens to show
  • Internal links that let a crawler reach every page from the home page

Enforcement

Checked, not hoped for

A standard nothing verifies is a preference. These run against the built output, which is what a visitor and a crawler actually receive, rather than against the source.

On every build
Titles, descriptions, canonicals, heading structure, alt text, image dimensions, inline script sanity. A build that would render a dead page fails rather than deploying.
On every publish
The same suite again, in quick mode, with anything unresolved named in the output rather than buried.
On demand
A full audit scored out of 100, with each finding written so it can be handed straight back as a fix.

One limit worth stating: the audit reads the built HTML, it does not open a browser. It catches the causes of layout breakage rather than measuring every symptom, and the report says so on its face instead of implying more confidence than it has.

Answers

Questions about this

Does moving content into a database hurt search visibility?

No, because the page is rendered twice. Every testimonial, question and answer is baked into the static HTML at build time, which is what a crawler receives, and the same list is refreshed in the browser so an edit shows up immediately for a visitor. The honest trade is that a visitor sees an edit at once and a search engine sees it at the next publish.

Why generate an llms.txt file?

It gives a language model a short, reliable description of the site instead of leaving it to infer one from whichever page it happened to read. It is generated from the site configuration and its key pages, so it cannot drift away from what the site actually contains.

Are AI crawlers blocked?

Search and citation crawlers are always allowed, because blocking them removes a site from AI answers while protecting nothing. Training crawlers are the one genuine choice, controlled by a single setting that flips the robots directives, the content signal and the llms.txt statement together so they cannot disagree.

Is the structured data checked, or just emitted?

Checked. The audit parses every JSON-LD block on every built page and fails on invalid JSON, warns on nodes with no type, and warns on empty values. Markup that describes content the page does not have is the thing search engines penalise, so the nodes are generated from the same data the page renders from.

See it on this site:llms.txt,robots.txt,the sitemap.