Free AI search audit
AI search audit: see your site the way AI crawlers and search engines do
Your pages look fine in a browser. AI crawlers and search engines may be getting something very different. This AI search audit tool loads every page it samples twice, as raw HTML and after JavaScript runs, then scores AI crawler access, LLM readiness and the technical SEO behind them: 24 factors in all. Every finding comes with a source.
What you get
A report you can hand straight to a client
Every finding tells you which page it’s about, what was seen there, and the source it rests on. Nothing in the report asks you to take our word for it.
One score, one clear verdict
You get a score out of 100 and a plain label: Healthy, Needs work, Critical issues or Blocked. Any caveat the audit has about its own number sits right beside it.
Six section scores
AI access, LLM content readiness, crawl and indexing, structured data, on-page SEO and performance. Each shows the weight it carries, so you can see where the points went.
Every finding, page by page
See how each factor performed on every page we sampled, worst first, with the evidence from that exact page.
A source behind every finding
We name the publisher and the document each finding relies on, and tell you whether that source is strong enough to justify a FAIL.
Fixes you can use
Each problem comes with what to do about it. For some, we generate the exact markup from your own page, ready to paste.
An honest list of what it couldn’t test
If the audit couldn’t observe something, it says so and gives the reason. That never counts against your site.
How it works
Four steps. About eight minutes.
Nothing gets sampled until robots.txt has been checked.
Check robots.txt first
The audit starts at the front door, the file that tells Google and AI crawlers what they may read. If robots.txt can’t be read, it stops and tells you, rather than calling the rest of your site healthy. A 404 is fine: Google treats it as no restrictions.
Pick a spread of pages
It collects the links on your homepage, removes anything robots.txt blocks, groups the rest by URL pattern and takes one page from each group. Think homepage, services, product, pricing, category, blog, author and about.
Fetch each page twice
Once as raw HTML, once after JavaScript has run. The gap between the two is the finding. It shows whether your content exists for crawlers that don’t run scripts, and many AI crawlers don’t. It’s a crawlability test for AI crawlers and Google in one pass.
Score, then cross-check
Page-level checks run first. Then come the cross-page checks: duplicate titles and H1s, canonical clusters and schema @id stability. If a slot in the sample can’t be filled, the report says why instead of padding it out.
How the score is built
24 factors. Six sections. Fixed weights.
The percentages below are the actual weights in your overall score, not guesses about what matters. The AI sections come first and make up 25% of the score. The other 75% is the technical SEO underneath, and AI search still depends on it: Google’s AI Overviews and AI Mode only link to pages that are indexed and eligible to show in Search with a snippet.
LLM / AI Access
Whether GPTBot, ClaudeBot, PerplexityBot and others are allowed in robots.txt, whether your content survives with JavaScript off, and whether you have an llms.txt. In other words, an AI crawler checker and an llms.txt checker in one section.
10%
LLM Content Readiness
Raw content availability, entity clarity, content structure, answer extractability and freshness. In short: can a model lift a clean answer out of your page?
15%
Crawl & Indexing
robots.txt, XML sitemaps, HTTP status codes, redirect chains, canonical tags and meta robots, plus whether each sampled page is eligible to be indexed at all.
30%
Structured & International
Structured data checked against a fixed schema set, and hreflang, including whether the language alternates you declare actually resolve.
20%
On-Page SEO
Title tags, meta descriptions, the H1 and heading order, and internal linking. Checked per page, then compared across pages for duplicates.
15%
Performance
Core Web Vitals from real-user field data where Google has it, with a lab test as the fallback. Reported page by page, never as one site-wide number.
10%
Reported, never scored
Two further factors appear in the report as advisory only. The evidence behind them isn’t strong enough to carry a score, so we show them without counting them.
Why trust it
Most free checkers guess. This one says when it doesn’t know.
A report you can send to a client is one that stays honest when the evidence runs out.
Every finding cites a source
Each factor carries a reference: the publisher, the document, and what it establishes for that finding. If the only backing is industry commentary, a FAIL is automatically capped at WARN, and the report tells you why.
It won’t score what it can’t test
A missing key, a blocked page, a browser-only check it couldn’t make: each is logged as NOT_TESTABLE with a reason. None of it is held against your site.
It won’t invent a rule
Where a factor has no published rules behind it, the audit shows it and declines to score it. No threshold is made up to fill a gap.
It tells you what it couldn’t see
Checks that need a real browser, like a collapsed accordion or an off-screen tab, report NOT_TESTABLE instead of guessing from markup.
Questions
What people ask before they run it
What it checks, what it won’t claim, and why a report sometimes says it couldn’t tell.
What is an AI search audit?
A check of whether AI search engines can reach, read and quote your site, and whether the technical SEO they rely on holds up. This one covers AI access (GPTBot, ClaudeBot, PerplexityBot, content without JavaScript, llms.txt) and LLM content readiness, plus crawl and indexing, structured data, on-page SEO and Core Web Vitals. It’s sometimes called an AI SEO audit or a technical GEO audit; you get a technical SEO audit in the same report.
Can AI crawlers access my site?
Run the audit and open the AI access section. For GPTBot, ClaudeBot, PerplexityBot and other AI crawlers it shows whether robots.txt lets them in. It also compares the raw HTML with the page after JavaScript runs, so you can see what’s left for a crawler that doesn’t run scripts.
Is my website AI ready?
For AI search readiness across a whole site, run this audit: it samples up to ten page types and scores AI access and LLM content readiness alongside the technical SEO underneath. To check one page in depth for ChatGPT, Perplexity, Claude and Google, use the AI Citation Readiness Checker; it takes about 30 seconds.
How long does it take?
About eight minutes for a ten-page audit. It fetches every page, runs a performance pass on each and scores 24 factors, so it’s doing real work the whole time. You can close the tab; the status link brings you back.
How many pages does it check?
Up to ten, picked to cover different page types rather than just the most-linked ones. Prefer to choose? Name up to ten specific pages instead.
Do I need to install anything?
No. It reads what’s publicly available, the way search and AI crawlers do. No tag to add, no access to grant.
Does it respect robots.txt?
Yes, strictly. Pages disallowed for Googlebot are removed from the sample before anything is fetched. Being able to reach a page isn’t the same as being allowed to.
What if my robots.txt is broken?
By default, the audit stops and reports BLOCKED rather than scoring the rest of your site as if nothing were wrong. You can choose to continue with the control-file checks instead.
Why does part of my report say NOT_TESTABLE?
Because something the audit needed wasn’t available: a key, a page it wasn’t allowed to fetch, or an observation it couldn’t make. It means “I couldn’t tell,” not “this is broken.”
Is this the same as the AI visibility report?
No. The AI visibility report asks what ChatGPT, Gemini, Google AI Overviews and AI Mode say about a brand. This audit asks whether your site is technically fit to be crawled, indexed and quoted in the first place. If the first result disappoints, run this one next.