Enter any page and see in 30 seconds whether ChatGPT, Claude, Perplexity and Google AI Overviews can crawl, read and cite it. Checks robots.txt, llms.txt, structure, readability and query coverage.
100% freeNo account30+ AI checks17 AI crawlers tested
Auditing
Fetching the page…
This usually takes 10–30 seconds. Keep this tab open.
These crawlers fetch pages in real time to answer questions and cite sources. Blocking them removes you from AI answers.
AI training crawlers
Used to build model knowledge. Blocking them is a legitimate choice, but it lowers the chance that models know your brand.
Readability & content density
Query coverage ·
Top 3 issues to fix first
Full report ·
Thanks! The full report is unlocked below. Adrian will email you a short personal review within one business day.
Free expert review
Unlock the full AI readiness report
Leave your email to see every check right now. Adrian, the SEO behind Marketer Coffee, will also send you a short personal note: what blocks AI citations on your page and what to fix first.
This audit checks one page. Our AI SEO (GEO) consultancy covers your whole site: crawler access, llms.txt, entity signals and content written to be cited.
Ranking first on Google is no longer the only thing that matters. More and more people ask ChatGPT, Perplexity, Claude or Google AI Overviews instead of scrolling through ten blue links. Those systems do not rank pages by backlinks. They fetch a page, strip it down to text, decide whether it answers the question and then quote one or two sources.
A standard SEO checker optimises for a crawler that indexes links. It will not tell you that your robots.txt blocks OAI-SearchBot, that there is no llms.txt, that your answer is buried under 400 words of intro, or that a sentence with 38 words is too dense for a model to summarise safely. If AI cannot read you, you are not just losing traffic. You are missing from the answer.
This free AI SEO audit tool simulates how an AI agent reads your page and returns one AI Readiness Score from 0 to 100, with the exact fixes that move it.
What this AI SEO checker analyses
Six areas, 30+ checks, one AI Readiness Score.
Every check is weighted. Crawler access and content structure count the most, because without them nothing else reaches the model.
01 / Access · 25 pts
AI crawler permissions
We parse your robots.txt the way Google does and test 17 agents: GPTBot, OAI-SearchBot, ClaudeBot, Claude-SearchBot, PerplexityBot, Google-Extended, Bingbot, CCBot, Applebot-Extended and more, for the exact path you entered.
02 / Discovery · 15 pts
llms.txt & indexing signals
Presence of llms.txt and llms-full.txt, HTTP status, noindex, nosnippet and max-snippet:0 directives, and the lang attribute that tells models which language they are reading.
03 / Structure · 20 pts
Structure for ingestion
One H1, no skipped heading levels, a main or article landmark, text-to-HTML ratio, lists and tables, and paragraph length. Models grab content in chunks, and a clean DOM gives them clean chunks.
04 / Readability · 15 pts
Readability vs complexity
Flesch Reading Ease, Gunning Fog index and average sentence length. Dense syntax raises the risk that a model misreads or hallucinates your message during summarisation.
05 / Coverage · 15 pts
Query coverage
We generate 20 probable user questions about your page topic (definition, cost, how-to, comparison, examples, risks and more) and test whether the content explicitly answers them. Low coverage means the page is thin for an answer engine.
06 / Trust · 10 pts
Entity & trust signals
JSON-LD structured data and useful types (Article, FAQPage, Organization, Person), a visible author, published or modified dates, and a complete title and meta description.
How to read your score
What your AI Readiness Score means.
A page that blocks every AI search bot or carries noindex is capped at 35–40 points regardless of content quality, because the content never reaches the model.
AI-ready80–100
Ready to be cited
Crawlers can reach the page and the content is structured to be quoted. Growth now comes from authority, mentions and topical depth.
Partly visible55–79
Visible, rarely cited
Models can read you but find gaps: missing llms.txt, thin answers or heavy sentences. Fixing the top three issues usually takes a day.
Invisible to AI0–54
Missing from AI answers
Blocked crawlers, noindex, JavaScript-only content or very thin text. Start with access, then structure.
FAQ
AI SEO audit: common questions.
What is an AI SEO audit?
An AI SEO audit checks whether large language models and AI search engines (ChatGPT, Perplexity, Claude, Gemini, Google AI Overviews) can crawl, understand and cite a page. It covers crawler permissions, llms.txt, content structure, readability, how many user questions the page answers and trust signals such as structured data and authorship.
Is this AI SEO audit tool really free?
Yes. Run as many audits as you like, with no account and no credit card. Email is only needed if you want to unlock the full check list and receive a short personal review.
What is llms.txt and do I need it?
llms.txt is a plain-text or Markdown file at the root of your domain that gives AI agents a clean summary of your site and links to its most important pages, without HTML noise. It is a young standard and not every model reads it yet, but it costs nothing and several AI crawlers already request it.
Should I allow GPTBot and ClaudeBot?
Search and answer bots such as OAI-SearchBot, Claude-SearchBot and PerplexityBot should be allowed if you want to appear in AI answers. Training crawlers such as GPTBot, ClaudeBot and Google-Extended are a business decision: blocking them protects your content from training, but models will know less about your brand.
How is the Query Coverage Ratio calculated?
The tool builds 20 question types that users typically ask about any topic (what it is, how it works, cost, steps, comparison, examples, risks, results and so on) and scans the page text for explicit answers to each. The ratio is answered questions out of 20. It is a heuristic, not a replacement for a real content audit.
Why is my page marked as thin or JavaScript-rendered?
Most AI crawlers do not execute JavaScript. If the raw HTML of your page contains under 150 words, the model sees an almost empty document, even if the rendered page looks rich. Server-side rendering or static HTML fixes this.
Does it audit my whole website?
No, it checks the one URL you enter, plus your domain-level robots.txt and llms.txt. Test your homepage, a key service page and your best blog post. A full-site AI visibility audit is part of our AI SEO consultancy.
What happens with my email?
Adrian uses it to send you one personal review of your result. No spam and no automated sequences. You can ask for your data to be deleted at any time.
We help brands become the source that ChatGPT, Perplexity and Google AI quote. Book a free call and we will tell you honestly what is worth fixing first.
AI crawler accessDiscoveryStructureReadabilityQuery coverageTrust signalsFetching the page…Reading robots.txt for 17 AI crawlers…Looking for llms.txt…Analysing heading structure and content density…Measuring readability…Testing 20 probable user questions…Calculating your AI Readiness Score…Please enter a valid page address, e.g. yourwebsite.com/blog/postWe could not load this pageThe page could not be fetched. Check that the address is correct, the page is public and it does not require a login or block automated requests. If it keeps failing, ask Adrian for a manual audit.We could not analyse this pageSomething went wrong while reading the HTML. Try again in a moment, or test a different page on the same site.Ready to be cited by AIAI crawlers can reach this page and the content is structured to be quoted. The next gains will come from authority, brand mentions and deeper topical coverage.Visible to AI, rarely citedModels can read this page but find gaps that lower the chance of a citation. Fix the top three issues below and re-test.This page is largely invisible to AIAccess, indexing or content problems are keeping this page out of AI answers. Start with the high-priority issues below.TestedHigh priorityMedium priorityPassedAllowedBlocked site-wideBlocked for this pathrobots.txt found and parsed for the exact path you entered.No robots.txt found: all crawlers are allowed by default.foundnot foundFlesch reading easeGunning FogWords per sentenceWords in HTMLText-to-HTML ratio20 questions a user might ask about “{topic}”, and whether the page text explicitly answers them.No significant issues found. This page is in good shape for AI search.{n} more issues and {p} passed checksA score of {n} usually means you are missing from AI answersLet’s go through your results together on a free 15-minute call. You will leave with a clear list of what to fix first, whether or not you work with us.Want to be cited by ChatGPT and Perplexity?This audit checks one page. Our AI SEO (GEO) consultancy covers your whole site: crawler access, llms.txt, entity signals and content written to be cited.All AI search and answer bots are allowedOAI-SearchBot, ChatGPT-User, Claude-SearchBot, Claude-User, PerplexityBot, Perplexity-User, Googlebot, Bingbot and DuckAssistBot can fetch this URL.{n} AI search bots are blocked: {list}These agents fetch pages live to answer user questions and cite sources. While they are blocked in robots.txt, your page cannot appear in those answers. Remove the Disallow rules for them.All AI training crawlers are allowedGPTBot, ClaudeBot, Google-Extended, Applebot-Extended, CCBot, meta-externalagent, Amazonbot and Bytespider can read this page, so future models will know your content.{n} AI training crawlers are blocked: {list}Blocking training crawlers is a legitimate choice, but models trained without your content are less likely to mention your brand. If visibility matters more than content protection, allow at least GPTBot and ClaudeBot.No robots.txt on this domainWithout a robots.txt every crawler is allowed, which is fine for AI visibility. Consider adding one anyway to point crawlers to your sitemap and llms.txt.llms.txt is presentAI agents can read a clean, Markdown-formatted summary of {host} without HTML noise.Both llms.txt and llms-full.txt are present, giving agents a summary and the full clean content of {host}.No llms.txt foundAdd a plain-text llms.txt at the root of {host} with a one-paragraph description of the site and a list of your most important pages. It takes 20 minutes and is already requested by several AI crawlers.Page returns HTTP 200The URL responds normally, so crawlers receive the content directly.Page returns HTTP {status}Crawlers expect a 200 response. Redirect chains, 4xx or 5xx codes waste the crawl and can drop the page from AI indexes.No noindex directiveThe page is allowed to be indexed by search engines and AI retrieval systems.Page carries a noindex directiveA noindex meta robots tag removes the page from Google, Bing and every AI system built on their indexes. Remove it unless the page is intentionally private.Snippets are allowedNo nosnippet, max-snippet:0 or noai directive, so AI systems may quote the page.Snippets are restrictednosnippet, max-snippet:0 or noai tells search engines and AI Overviews not to quote your content. Remove the directive if you want to be cited.Language declared ({lang})The html lang attribute tells models which language the content is in.No html lang attributeAdd lang to the html element, e.g. lang=”en”, so models and search engines do not have to guess the language.Exactly one H1A single H1 gives models an unambiguous topic for the page.{n} H1 headings foundMultiple H1s blur the main topic. Keep one H1 and turn the others into H2.No H1 headingWithout an H1 the model has to infer the topic from the title and body. Add one H1 that states what the page is about.Heading hierarchy is consistentNo skipped levels (for example H2 to H4), so the outline maps cleanly to sections.{n} skipped heading levelsJumping from H2 to H4 breaks the outline that models use to chunk content. Use consecutive levels.Content is split into {n} H2 sectionsSections let an answer engine grab exactly the part that answers a question.Too few H2 sections ({n})Long content without H2 headings is one undifferentiated block for a model. Add descriptive H2s, ideally phrased as the questions users ask.Main content landmark presentA main or article element tells parsers where the primary content starts and ends.No main or article landmarkWrap the primary content in a main or article element so extractors can separate it from navigation and footer.Healthy text-to-HTML ratio ({ratio}%)Enough visible text relative to markup; the page is cheap for a model to ingest.Low text-to-HTML ratio ({ratio}%, {scripts} scripts)Most of the document is markup or scripts. Reduce boilerplate, inline SVGs and tracking code, and make sure the main text is in the raw HTML.Lists or tables present ({n})Lists and tables are the easiest structures for models to quote verbatim.No lists or tablesTurn steps, comparisons and features into bullet lists or tables. AI answers reuse these structures directly.Paragraphs are short enoughNo paragraph exceeds 90 words, which keeps each idea in a single chunk.{n} paragraphs over 90 wordsLong paragraphs mix several ideas in one chunk. Split them so each paragraph makes one point.Flesch reading ease {v} (readable)The text is easy enough for a model to summarise accurately. Calibrated for English; other languages score slightly lower.Flesch reading ease {v} (dense)Dense prose increases the risk that a model misreads or hallucinates your point. Shorter words and sentences raise the score. Calibrated for English; other languages score slightly lower.Gunning Fog {v}Roughly the reading level of a general audience; {pct}% of words are complex.Gunning Fog {v} (hard to read){pct}% of words have three or more syllables. Replace jargon with plain words where meaning allows.{v} words per sentence on averageSentences are short enough to be quoted as standalone facts.{v} words per sentence on average (too long)Aim for under 20 words per sentence. A model quotes single sentences, and long ones rarely survive as clean answers.Answers {n} of 20 probable user questionsThe page explicitly covers most intents a user could have about this topic, which makes it a safe source for an answer engine.Answers only {n} of 20 probable user questionsThe content is semantically thin for an answer engine. Add short, explicit sections for the unanswered questions listed in the Query coverage card.Answers come first under {pct}% of headingsMost sections open with a direct answer of 60 words or fewer, exactly what AI snippets extract.Only {pct}% of sections open with a direct answerStart each section with a one- or two-sentence answer, then elaborate. Models extract the first chunk under a heading.FAQ or question-style headings presentQuestion headings or FAQPage schema map directly to how users phrase prompts.No FAQ or question-style headings ({n})Add two to five headings phrased as real user questions, or a short FAQ with FAQPage schema.Structured data present ({types})JSON-LD gives models typed facts about the page instead of guesses.No JSON-LD structured dataAdd schema.org JSON-LD describing the page type, organisation and author. Rank Math or Yoast can generate it.Useful schema types ({types})Types such as Article, FAQPage, Organization and Person help models connect the page to an entity.No content or entity schema typeStructured data exists but has no Article, FAQPage, Organization, Person, Product or Service type. Add the one that matches the page.Author signal foundA visible author or Person schema supports experience and expertise signals.No author signalAdd a byline with a real person and author markup. Models weigh who wrote a page when choosing sources.Published or modified date foundDates let models judge freshness, which matters for fast-moving topics.No published or modified dateAdd datePublished and dateModified in schema or a visible time element. Undated content is treated as potentially stale.Title and meta description completeTitle ({tl} characters) and description ({dl} characters) give retrieval systems a clear summary.Title or meta description missing or shortTitle has {tl} characters and the description {dl}. Write a descriptive title and a 120–160 character description; both are used in retrieval.Very little text in the raw HTMLOnly {n} words were found in the HTML source. AI crawlers see an almost empty page. Add real content or make sure it is rendered server-side.Only {n} words were found in the HTML source but the page loads many scripts, so the content is probably rendered by JavaScript. Most AI crawlers do not execute JavaScript. Use server-side rendering or static HTML.What is {topic}?How does {topic} work?How do I do it, step by step?How much does it cost?What are the benefits?What are the drawbacks or limitations?How does it compare to alternatives?Which option is best or recommended?Are there examples or case studies?Who is it for?How long does it take?What mistakes or problems should I avoid?Is there data or research behind it?Which tools or software are involved?What do I need before starting?Is there an FAQ or question-based section?What is current or changing in 2026?What are the risks, legal or safety aspects?What results can I expect?What is the next step or how do I get started?
To provide the best experiences, we use technologies like cookies to store and/or access device information. Consenting to these technologies will allow us to process data such as browsing behavior or unique IDs on this site. Not consenting or withdrawing consent, may adversely affect certain features and functions.
Functional
Always active
The technical storage or access is strictly necessary for the legitimate purpose of enabling the use of a specific service explicitly requested by the subscriber or user, or for the sole purpose of carrying out the transmission of a communication over an electronic communications network.
Preferences
The technical storage or access is necessary for the legitimate purpose of storing preferences that are not requested by the subscriber or user.
Statistics
The technical storage or access that is used exclusively for statistical purposes.The technical storage or access that is used exclusively for anonymous statistical purposes. Without a subpoena, voluntary compliance on the part of your Internet Service Provider, or additional records from a third party, information stored or retrieved for this purpose alone cannot usually be used to identify you.
Marketing
The technical storage or access is required to create user profiles to send advertising, or to track the user on a website or across several websites for similar marketing purposes.
