Every check, all 82 in plain language

Every page has a title, at a length that survives the results page, saying what the page actually is.

Each page ships its own description, so the search engine isn't left to write your snippet for you.

One clear H1 per page and no skipped levels, the outline crawlers read to understand your structure.

Every page declares its preferred URL, so duplicates don't split ranking authority between them.

No stray noindex or nofollow quietly hiding pages you spent months building.

Links that hop through two or three redirects before landing. Every hop leaks authority and speed.

Readable, lowercase, meaningful paths instead of query-string soup.

Whether a bot can genuinely walk from your homepage to this page without hitting a wall.

Which pages are eligible to appear at all, after robots.txt, meta robots and canonicals are resolved together.

Every page has inbound links from elsewhere on the site. Orphan pages don't rank.

How many clicks from the homepage each page sits. Past three or four, crawlers stop coming back.

Your important pages are in the XML sitemap, and the sitemap isn't listing pages you've blocked.

404s, 500s and soft errors reached from your own links, with the exact page that links to each.

The routes a bot takes through your site, and precisely where they dead-end.

Session IDs, tracking parameters and mixed casing, the quiet factory for duplicate URLs.

Paginated series linked so crawlers can walk them and don't treat every page as a duplicate.

One canonical form for trailing slashes, instead of two live URLs for every page on the site.

The whole graph, rendered as a tree, a force graph, or a sortable table, hubs, clusters and isolated pockets.

Inbound and outbound links per URL, so you can see what each page supports and what supports it.

The depth distribution across your site, and what is buried too deep to earn traffic.

Language and region variants declared, and declared reciprocally, which is where most setups break.

A declared fallback for visitors whose language you don't serve.

The language you declare matches the language actually on the page.

A lang attribute on the html element, read by screen readers, translators and crawlers alike.

The emerging standard that tells AI crawlers what your site is and where the substance lives.

The expanded manifest: your key content in one clean, model-readable file.

Whether GPTBot, ClaudeBot, PerplexityBot and their peers are allowed in. Many sites block them by accident.

Your robots.txt and meta rules for AI agents, read back to you explicitly instead of assumed.

Whether your content survives without JavaScript. If a model can't read it, you don't exist to it.

Clear, self-contained sections a model can quote without dragging in half the page.

The answer near the top, ahead of the preamble, how models decide what is worth lifting.

Questions marked up as questions, so the answer beneath them is attributable to you.

Semantic HTML (real headings, lists and tables) instead of a thousand nested divs.

Pages that answer questions but carry no FAQ markup, the cheapest snippet win on your site.

Procedural content that could be a step-by-step result, sitting unmarked.

A concise, standalone answer an engine can lift verbatim without editing.

Headings phrased the way people actually search, rather than the way your org chart talks.

Passages short and clean enough for a voice assistant to read out loud.

Lists, tables and definitions, the shapes answer boxes reach for before prose.

A composite read on whether this page could realistically take position zero.