Every check we run

All 191 of them, grouped the way the report presents them. Every finding carries the evidence from your page and the markup that fixes it, on every plan including the free one.

Meta tags and the search snippet 5 checks

What the result looks like in search, measured the way search engines measure it.

Title, presence, character length and the rendered pixel width for desktop and mobile, which is what actually decides truncation.
Meta description, presence, length, and whether it will be cut off.
SERP preview, the result drawn at real widths rather than described.

Crawling and indexing 52 checks

The checks that decide whether a page can be found at all. A failure here makes every other check academic.

HTTP status, non-200 responses that stop a page being indexed.
Redirects, hop counts, loops, and HTTPS to HTTP downgrades.
Robots directives, noindex, nofollow and noarchive from the meta tag and the X-Robots-Tag header, which markup-only tools miss.
robots.txt, presence, whether it blocks everything, and whether it points at your sitemap.
Sitemap, presence, valid XML, and whether robots.txt declares it.
Canonical, missing, cross-domain, protocol mismatch, or not self-referential.
hreflang, invalid codes, duplicates, and missing self-reference.
URL shape, length, depth, parameter count, encoded characters, underscores, uppercase, and whether it parses at all.
Language, a valid lang attribute.
Viewport and charset, declared, and declared correctly.
Mixed content, split into what browsers block outright and what only breaks the padlock.
Favicon, declared and reachable.

Content and metadata 24 checks

Whether the page says what it is about, in a shape both readers and machines can follow.

Heading topology, a single H1, no skipped levels, no empty headings.
Content depth, word count against what the page is trying to do, and the ratio of text to markup, which is what separates a thin page from a heavy one.
Keywords, placement and distribution, and stuffing in either direction.
Open Graph and Twitter cards, completeness and card type.
Delivery, total page weight, cache headers, and how much of it belongs to third parties rather than to you.

AI and answer readiness 40 checks

Whether an assistant can quote this page, and whether you have said it may. This is the newest part of the audit and the part most tools do not run at all.

AI crawler policy, GPTBot, ChatGPT-User, OAI-SearchBot, ClaudeBot, Claude-Web, Google-Extended, PerplexityBot, Applebot-Extended, CCBot, Bytespider and more, each read from your robots.txt and reported as allowed or blocked rather than guessed at.
llms.txt, presence and validity of the emerging convention for stating terms to model crawlers.
Answer readiness, whether the page answers a question in a form that can be extracted: scannable structure, named attribution, and answer-shaped paragraphs.
Stale content, undated pages, and dates old enough that an assistant will prefer somebody else.
Entity extraction, whether the subject and publisher are named clearly enough to be identified.
Hydration, what exists only after JavaScript runs. A title, body or schema that is client-only is invisible to anything that does not execute the page.

Speed and delivery 18 checks

Core Web Vitals from Google's own field data where it exists, and the delivery problems behind them.

LCP, INP, CLS, real-user data from the Chrome UX Report, with the lab figures beside them and the difference explained.
Time to first byte, how long the server takes before anything can start, which no amount of front-end work can recover.
Render blocking, stylesheets and scripts in the way of the first paint.
Compression, whether text is served compressed.
HTML size, documents large enough to delay first paint, and bodies so long they are truncated.
Request count, how many resources the page pulls before it is usable.
Deprecated markup and inline styles, what is adding weight for nothing.

Security 18 checks

The headers and certificate a browser uses to decide whether to trust the page. We send every one of these ourselves, which you are welcome to verify by running this on us.

TLS certificate, validity, and expiry that is close enough to matter.
HSTS, presence, and whether the max-age is long enough to be useful.
Content-Security-Policy, presence, and whether it is undermined by unsafe-inline.
Clickjacking, whether anything stops the page being framed.
X-Content-Type-Options, whether browsers may guess a resource type and execute it.
Referrer-Policy, whether full URLs leak to third parties.
Server version disclosure, software versions announced in headers.

Links, media and schema 34 checks

The parts of the page that point somewhere else, and the markup that explains it to a machine.

Broken links, every link followed and reported with its status.
Anchor text, empty anchors, text that says nothing, and the same words pointing at different places.
Undisclosed affiliate links, outbound links that look commercial and carry no rel attribute saying so, which is a disclosure problem before it is an SEO one.
Link distribution, internal against external, and excessive outbound.
Dead fragments, links to an anchor that does not exist on the target.
Alt text, missing, and stuffed.
Image dimensions, declared, so the layout does not shift.
Image formats, legacy formats where a modern one would do, and images served at one size for every screen.
Lazy loading, used, and not used on the hero image, which delays LCP.
Schema, valid JSON-LD, required properties present, and whether it is eligible for a rich result.