AI Insights

Stop stacking SEO addons: Statamic's open-source fix for search and AI visibility

On 9 October 2026, Belgian Statamic agency Vulpo published something UK teams on Statamic have needed for a while: one open-source addon that replaces the usual tangle of meta tags, sitemaps, redirects and half-finished "GEO" experiments — and quietly answers the question almost nobody can: are AI assistants actually reading this site?

That question sits next to Google's own advice. In its guide to generative AI features on Search, Google is blunt: for AI Overviews and AI Mode, foundational SEO still wins. Fancy files and "GEO hacks" do not. The useful work is clearer content, crawlable pages, and measurement — not another plugin that promises citations.

The Statamic SEO mess Vulpo kept finding

Vulpo's GEO audits of Statamic sites kept landing on the same pattern. SEO was rarely one system. It was a meta-tag addon, a separate sitemap package, a redirect tool, sometimes a hand-rolled JSON-LD partial in the layout, and occasionally a GEO addon on top. Each had its own settings screen and its own idea of what the page title should be. When two disagreed, the page shipped both answers.

And almost nobody could say whether ChatGPT, Claude, Perplexity or Gemini had ever requested the site. Plenty of teams had added an llms.txt. Very few had a log.

That is not a Statamic failure. It is what happens when "SEO" is treated as a shopping list of addons instead of one coherent publish path — titles, canonicals, redirects when a slug changes, structured data that matches what the page actually says, and a way to see crawler behaviour without guesswork.

What SEO & GEO actually ships

Vulpo's SEO & GEO addon (MIT, Packagist and GitHub) is aimed at that whole stack in one Control Panel section and one template tag. The highlights that matter for UK SMEs:

For editors

  • Live Google and social-card previews as you type, with titles truncated by measured pixel width (how Google cuts them), not a blunt character count
  • One place for title, description, canonical, robots, Open Graph, Twitter cards, hreflang and Search Console verification — page overrides site defaults; multisite can override again

For search engines

  • /sitemap.xml from published, routable entries, with exclusions and automatic sitemap-index behaviour when the URL list grows
  • Redirects (exact, wildcard, regex; 301/302/410) checked only when a URL would otherwise 404 — so normal page loads pay nothing
  • Automatic redirects when a URL changes (renamed slug, structure move, date change) — the quiet failure Vulpo saw most often in audits
  • A 404 log per site, with hit counts and referrers, and one-click promote-to-redirect

For AI assistants (honestly framed)

  • Structured data: Organization or LocalBusiness, WebSite, BreadcrumbList, plus per-page FAQ, Article, Service, Person, Product or Event
  • /llms.txt with an editable site summary
  • robots.txt with an AI crawler policy you can set in the CP
  • An AI crawler log — which assistants requested the site, surfaced beside recent 404s

Requirements are straightforward: PHP 8.3 and Statamic 6. Install is two commands, then {{ vulpo_seo }} in the layout. Flat-file by default (settings and redirects in YAML under version control); Eloquent-driven projects move the addon's data to the database without extra config. Migration from SEO Pro or Vulpo's earlier GEO addon includes a dry-run path so you are not guessing on go-live day.

What it deliberately does not promise

Vulpo is clear, and UK buyers should copy that clarity into any vendor conversation: no addon can get your site cited by an AI assistant. Structured data and llms.txt can make a site easier to understand. Whether a given system then mentions your business is decided elsewhere.

That is why the crawler log matters more than the buzzwords. It does not promise an outcome. It replaces guesswork with observation: which assistants visit, how often, and whether that changes after you improve content. Readiness and measurement are what can honestly be delivered.

Google's Search Central guide lands in the same place from the other direction. For Google Search's generative features:

  • Meet ordinary Search technical requirements (indexed, snippet-eligible)
  • Create unique, useful, people-first content — not commodity "seven tips" pages
  • Keep a clear technical structure and good page experience
  • You do not need llms.txt, special AI markup, content "chunking", or inauthentic mention-farming for Google's generative features — Google Search ignores those files for ranking and visibility
  • Structured data is still useful for rich results overall, but there is no special schema required purely for generative AI features
  • Measure with Search Console's Generative AI performance report when it is available to your property

So the practical UK brief is not "do GEO instead of SEO". It is: do SEO properly, make AI readiness observable, and ignore theatre.

Statamic's wider October signal: AI that respects the CMS

The same fortnight underlines why AI and Statamic fit when the tools respect the model.

Ghostwriter (1994) shipped its 1.0 on 2 October 2026 and moved fast through 1.2 by 3 October: drafts that learn the site's voice from existing entries, never invent figures or quotes you did not give, start unpublished, and block publish while "finish this page" gaps remain. OpenRouter landed as a fourth provider. That is the opposite of paste-and-hope chatbots — it is AI constrained by blueprints, permissions and human review.

Statamic 6.35.1 (6 October 2026) was a security release: hardened asset authorisation, tighter SVG sanitisation, remote-URL validation and safer user file paths. Unsexy, necessary. If you are adding SEO, crawler logs or AI drafting, keep the core on a current patch.

Together with earlier MCP and blueprint-aware tooling, the pattern for UK teams is consistent: structure first, agents second. Statamic's blueprints, flat files and Git-friendly content give assistants something real to work with. Visibility addons and drafting tools work best when they sit on that structure — not when they invent a parallel SEO universe in four plugins that disagree about the title tag.

A practical checklist for UK Statamic teams this quarter

  1. Audit the SEO stack. Count how many addons (or partials) currently set titles, robots, sitemaps and redirects. If the number is "more than one", you have a conflict risk.
  2. Fix URL change behaviour. Confirm that renamed or moved pages create redirects automatically — or schedule a one-off redirect import from analytics 404s.
  3. Separate readiness from citation. Ship crawlable pages, sensible schema that matches visible content, and a crawler policy you can explain. Do not buy a promise of "AI mentions".
  4. Measure two channels. Search Console (including generative AI reporting when available) for Google; an AI crawler log (or server logs) for assistants outside Google.
  5. Keep content non-commodity. Google's own guide puts unique, expert, people-first pages above markup tricks. A cleaner addon cannot rescue thin service pages.
  6. Patch Statamic, then add. Land on a current 6.x security release before layering SEO or AI writing tools.
  7. If you use AI drafting, demand publish guards. Prefer tools that refuse to invent facts and keep drafts unpublished until a human finishes the gaps.

Why this matters for Clarity Growth

Clarity Growth partners with Statamic and builds AI-ready content systems for UK SMEs — including CMSless work where structured modelling and honest tooling beat page-builder chaos. Vulpo's 9 October release is useful partner material because it attacks a real operational pain (fragmented SEO) and pairs it with an honest AI-visibility story that matches Google's own guidance. We help clients choose and implement a coherent search stack on Statamic, wire redirects and structured data into the publish workflow, and connect measurement (Search Console, crawler behaviour, content quality) so "AI visibility" stays a commercial discipline rather than a buzzword budget line.

Sources