Running your first scan
What AIRRNK checks, how long it takes, and how to read the report.
title: "Running your first scan" slug: "first-scan" description: "What AIRRNK checks, how long it takes, and how to read the report." group: "Getting Started" order: 2
A scan is the full 47-point AI-readiness audit plus a fresh citation sweep across four answer engines. It runs live — no queuing — and completes in 30 to 60 seconds for most sites.
Starting a scan
There are three ways to trigger one:
- Home page input. Paste any URL into the scanner on airank.tech. No signup. No credit card. Full report.
- Dashboard. Connected sites auto-scan every six hours. You can also manually re-scan from the site's settings page.
- API.
POST /api/scanswith a{ site_id }body. See the API reference.
What gets checked
The scan runs in three parallel tracks.
Track 1 — Technical readiness (17 checks)
Crawl access (robots.txt, llms.txt, sitemap), status codes, canonical tags, hreflang, structured data validity, <head> hygiene, internal link graph, page weight, Core Web Vitals, server response times, and mobile parity. The usual suspects, biased toward what AI crawlers specifically respect.
Track 2 — Content extractability (18 checks)
For each of your top pages, we measure: chunk density (how many quotable 100–400 token spans per page), claim uniqueness (are your sentences findable anywhere else on the web?), schema coverage (FAQPage, Product, HowTo, Article), heading hierarchy, author attribution, citation-worthy formatting (lists, tables, blockquotes that survive chunking), and freshness signals.
Track 3 — Citation telemetry (12 checks)
Your query list runs against ChatGPT (browsing on), Claude (web search tool enabled), Perplexity (default model), and Google AI Mode. We record: citation count, snippet extracted, competitor citations on the same query, and snippet sentiment.
Reading the report
The report has five sections, top to bottom.
AI Score
A single number out of 100. Anything above 70 is good. Above 85 puts you in the top decile of sites we scan. Under 50 means you have easy wins — fixes in the 2–5 point range are sitting there.
Citation leaderboard
For each of your tracked queries, we show who got cited and in what order. If a competitor shows up, we mark it. If you show up, we show the snippet and source URL.
Highest-lift fixes
The top five. Each fix includes:
- A plain-English explanation of the issue.
- The expected AI Score delta if you ship the fix.
- The estimated effort (minutes to days).
- A one-click "apply" button on paid plans (writes directly to your site when the plugin/app is installed).
Full 47-point breakdown
Expandable. Every check, every status (pass/warn/fail), every fix hint. This is where engineers live.
Change log
Every delta since your previous scan, so you can confirm fixes shipped and see regressions immediately.
How long between scans?
| Plan | Scan frequency |
|---|---|
| Free | 24 hours |
| Pro | 6 hours |
| Team | 1 hour |
| Enterprise | 15 minutes (on-demand) |
Citation sweeps follow the same cadence, with an overnight full run that uses a larger query pool.
What we don't scan
- JavaScript-rendered content that requires auth to see. The scanner uses the same crawlers AI platforms use — mostly Bingbot-class user-agents — and those cannot log in.
- Pages behind a paywall.
- Content in PDFs, unless you explicitly opt in (they are slow to process).
If your product lives behind a login, see the docs pattern for exposing public reference material for AI crawlers without compromising the gated experience.