TL;DR

Most "full" SEO audits don't actually cross-reference your crawl data with your authority data, so they can't tell you whether a broken canonical on a high-traffic page is a crisis or a non-issue. nqzai's full audit does exactly that, but only if you already have both an on-page crawl and an off-page authority report on file—it never bills you for a synthesis it can't perform. The on-page crawl checks roughly three dozen signals per page, including a hard flag for pages over 3MB and a unique check for AI-agent readiness (robots.txt, sitemap.xml, llms.txt, agent.json, OpenAPI spec, and WebMCP headers). The off-page half measures domain authority, total backlinks, referring domains, dofollow percentage, and top-ranking keywords broken into position bands (1, 2–3, 4–10), plus anchor-text analysis for toxic patterns like gambling or pharma links.

The bottom line: if you want a verdict that tells you which technical defects are actually costing you traffic versus which are noise, run the on-page and off-page audits first, then ask for the full audit—otherwise you'll just get a polite refusal and a prompt to run the missing pieces.

Run a Full SEO Audit

Most "SEO audit" tools do one thing well: they crawl your pages, flag a pile of issues, and hand you a score. That's useful, but it's half the picture — a site can have a spotless crawl and still not rank because it has no authority, or it can have plenty of authority and be squandering it on pages full of broken canonicals and duplicate titles. Neither a crawl report nor a backlink report can tell you which problem you actually have, because each only sees its own half. nqzai's full SEO audit is built to answer that specific question, as a chat capability — you ask for it in plain language and get a diagnosis, not just a checklist.

How you run it

Inside the nqzai chat, any of these trigger it directly, with no extra clicks:

  • "Run a full SEO audit"
  • "Audit my site"
  • "Run every SEO audit"

The system recognizes "full" or "complete" SEO audit phrasing as an explicit, high-confidence request and routes straight to the audit rather than asking a clarifying question first — it's built to skip the back-and-forth when you've already been specific. If you just say "check my SEO" without qualifying it, you may be offered a choice between the full audit and its two components (technical on-page, off-page authority) so you're not billed for more than you asked for.

What "full" actually means

Direct answer: Here's the part that's easy to assume wrong: the full audit itself does not go crawl your site or buy new backlink data. It reads what's already been measured about your domain and reasons across it. Specifically, it pulls together up to six sources on file for your site:

  • On-page technical crawl — the page-by-page scan of your site's HTML, structure, and status codes.
  • Off-page authority and rankings — domain authority, backlink profile, and the keywords you currently rank for.
  • Analytics traffic — which pages are actually earning visits, from your connected Google Analytics.
  • Content quality — scored assessments of your published pages.
  • AI-answer visibility — how often AI assistants and AI Overviews mention or cite you.
  • Your tracked keywords — the terms you've chosen to watch, whether or not you rank for them yet.

If the on-page and off-page pieces haven't been run yet, the full audit says so plainly and tells you which two to run first — it never charges you for a synthesis it can't actually perform. Once at least one of those two exists, it produces a report. It's built this way rather than re-running a fresh crawl every time because your on-page and off-page data already exist as their own separately-priced capabilities, and stapling two reports side by side under one heading isn't actually a third report — it's the same information twice. The value of a full audit is what only becomes visible when both halves are read together, and that's what the rest of this capability does.

What the on-page crawl checks

Direct answer: When you run the on-page piece (on its own, or as a prerequisite to the full audit), the crawl scores your site against roughly three dozen individual checks, grouped into:

  • Titles & meta — missing or duplicate <title> tags, titles over 60 characters, missing or duplicate meta descriptions, titles or descriptions that don't match the page's actual content.
  • Headings — missing H1, multiple H1 tags on one page.
  • Content quality signals — thin content (low word count), low readability, oversized pages (including a hard flag past 3MB), and leftover placeholder ("lorem ipsum") text that shipped by accident.
  • Images — missing alt text, missing title attributes.
  • Canonical & indexation — canonicals pointing at broken pages, canonicals pointing at redirects, recursive canonical loops, chained canonicals.
  • Links & redirects — redirect chains, links pointing at redirects, broken outgoing links, broken images/CSS/JS resources, orphan pages with no internal links pointing at them.
  • Status & performance — 4xx/5xx pages, render-blocking resources, slow page loads, slow server response time.
  • Technical hygiene — missing doctype, no compression, missing favicon, deprecated HTML, Flash, frame-based content.

On top of the per-page checks, the crawl runs site-wide relational checks that no single page can reveal on its own — duplicate titles, duplicate meta descriptions, or duplicate body content shared across multiple pages; whether HTTP actually redirects to HTTPS; whether both the www and non-www versions of your domain resolve without redirecting to one canonical host; and whether a made-up URL correctly returns a 404 instead of a soft-200. If your SSL certificate is nearing expiry, that gets flagged too, with the actual expiry date rather than a vague warning.

There's also a check most audit tools skip entirely: AI-agent readiness — whether your robots.txt, sitemap.xml, and llms.txt are present, and whether you expose newer machine-readability surfaces like an agent.json, an OpenAPI spec, or a WebMCP header. This matters increasingly for how AI crawlers and browsing agents, not just Googlebot, read and act on your site.

When your Google account is connected, the on-page report also cross-references Search Console index coverage against the pages it crawled — so it can tell you not just what's technically broken, but which of your live pages Google hasn't indexed at all, and (via impressions) which pages carrying issues are actually the ones earning search visibility. A duplicate-title defect on a page nobody sees is a very different priority than the same defect on a page earning most of your impressions.

What the off-page check measures

Direct answer: The off-page half looks at your authority and ranking footprint: domain authority, total backlinks, referring domains, the percentage of links that are dofollow, and your top ranking keywords broken into position bands (position 1, positions 2–3, positions 4–10). It also runs anchor-text analysis — flagging toxic anchor patterns (gambling, pharma, and similar link-scheme language), over-optimized keyword-stuffed anchors past the ~60% threshold that reads as unnatural, and a link-health pass that checks whether the pages your backlinks actually point at are still alive.

The part neither half can see alone

Direct answer: This is where the "full" in full audit earns its name. Once both halves are on file, the audit runs a set of joins across them — findings that are only visible when a fact from the crawl and a fact from the authority check are read against each other:

  • A backlink lands on a dead page. An external site linked to you, and the crawl shows that URL now returns a 404 or 500. The authority those links carry is being thrown away, and the fix is a redirect, not a link-building campaign.
  • A backlink lands on a page Google hasn't indexed. The link is real, the page is alive, but it's invisible in search — so nothing the link earned can be spent.
  • A page that already ranks is carrying defects. If a URL ranking on page one has a blocking technical fault, that's flagged as the single highest-priority fix on the whole site — the position is already earned and at risk, which is a different urgency than an issue on a page nobody finds.
  • A page earning real traffic (from your connected Analytics) has crawl issues. Real visitors are meeting the broken version right now.
  • Descriptive anchor text and your actual rankings point in different directions — worth knowing even though it isn't something to "fix" outright.
  • A ranking page scores poorly on content quality, or you're tracking keywords you don't rank for at all.

From there, the audit names the actual constraint: whether your bottleneck is on-page (you have authority you're not converting), off-page (your pages are clean but you have nothing to rank them with), both, or — on a healthy site — neither, in which case the remaining gains are page-specific rather than structural. That verdict is computed from your measured scores, not guessed by a model.

Findings drawn from a source older than 30 days are shown, never hidden, but automatically downgraded from "go fix this" to "worth checking" — a defect measured a month ago may already be resolved, and the audit won't tell you to fix something with outdated confidence.

Illustrative example (not a real customer report)

Direct answer: The figures below are constructed to show the kind of output the audit produces — they are not from an actual audit.

Site: example.com — On-page 71/100, Off-page 43/100 (domain authority 18/100) Constraint: Off-page. "Pages are in good technical shape (71/100) and the site still has little authority to rank with (18/100). More on-page work will not move this — earning referring domains will." Top finding: A ranking page (position #6 for "wireless charging pad") carries a blocking fault — noindex meta tag left on by mistake. Fix this first; the position is already earned. Also flagged: 3 external links point at a page returning a 404. 12 pages share a duplicate meta description.

What it costs

Direct answer: The synthesis step itself is inexpensive — it makes no new provider calls, so there's no separate approval gate for it. The on-page and off-page audits it reads from are priced and approved separately, shown in tokens (never dollars) before you run them, scaled to your site's page count and backlink volume.

Try it

Direct answer: In chat, just say "run a full SEO audit." If you haven't audited your site before, nqzai will tell you exactly which two audits to run first — and once they're on file, ask again and you'll get the diagnosis, not just two reports stapled together.

Evidence, limits, and reproducible use

Direct answer: Reproducible workflow. Supply the exact site, property, date range, and comparison period. Review the source data, filters, and assumptions before using a finding to change content or reporting. Record the run date because search data is revised and delayed.

Limit. The output is diagnostic evidence, not a ranking guarantee. Search performance depends on crawling, indexing, competing pages, and user demand that a single report cannot control.

For the currently exposed nqzai workflow and connection limits, check the public capabilities inventory before relying on a result.

Primary references