Readable structure 25 points
Title, meta description, a single H1, heading hierarchy with no skipped level, canonical URL, XML sitemap, robots.txt. What every engine reads before deciding whether to keep the page.
The score is computed on your page, live, and shown before you give us a single e-mail address. It takes about a minute.
What a machine cannot do on its own : ask ChatGPT, Claude, Perplexity and Gemini about your market to see who they cite in your place, read your templates one by one, and cost the work. We do it by hand and send it within 48 working hours.
We will come back to you within 48 working hours with the full report and the citation record.
Title, meta description, a single H1, heading hierarchy with no skipped level, canonical URL, XML sitemap, robots.txt. What every engine reads before deciding whether to keep the page.
JSON-LD structured data and markup validity, declared entity, question-and-answer passages, dated freshness, access granted to GPTBot, ClaudeBot, PerplexityBot, OAI-SearchBot and Google-Extended, presence of an llms.txt.
Text alternatives, form labels, link text that makes sense out of context, accessible names for buttons, frame titles, table headers, declared language. WCAG 2.1 AA and RGAA criteria that can be checked on the served code.
HTTPS, redirect chain, document weight, compression, and performance measured through Google's PageSpeed Insights API — the same engine that feeds the Core Web Vitals signals.
Want to understand it first ? Our page on GEO explains how a model picks the sources it cites.
The diagnostic reads the code your server actually serves. It guesses nothing : every point gained or lost matches a finding you can verify yourself. Colour contrast and keyboard navigation are not measured automatically — they need a rendered page and a human test, and they are part of the full audit.
Yes, and without an e-mail. The score and the findings are computed on your page the moment you start the analysis, then displayed on screen. You only give us your details if you want the full report and the citation test on generative engines, which take human work.
Generative Engine Optimization means making content selectable and citable by generative engines — ChatGPT, Claude, Perplexity, Gemini, Copilot — rather than only rankable by Google. The levers differ: declared entities, extractable passages, dated freshness, robots granted access.
No. A generative engine leans heavily on pages the classic index has already found and judged reliable. GEO adds a layer: it makes content liftable once it has been found. The two are worked together, and our diagnostic measures both.
Because the mechanisms overlap. A properly ranked heading, a described image, a labelled field, a declared language: these are the same signals that make a page intelligible to a screen reader and to a language model. In Europe, accessibility is also a legal obligation for a range of organisations.
No citation is possible. Many sites block GPTBot, ClaudeBot or PerplexityBot without knowing it, inherited from a configuration or a plugin. The diagnostic reads your robots.txt and tells you which ones are shut out.
The free diagnostic covers the address you enter, plus the domain's configuration files — robots.txt, sitemap.xml, llms.txt. The full audit covers every template on the site and the citation test across engines.
Describe what you need. We turn an institutional requirement into a solution that is reliable, secure and built to last.