16 free tools that check whether AI crawlers can reach, read and quote your website. No signup, no credit card and no AI model behind the scores.
AI search works differently from classic search. ChatGPT, Claude, Perplexity, Gemini and Copilot send their own crawlers, which fetch the HTML of a page and, in most cases, stop there. They do not run your JavaScript. They use separate crawlers to collect training data, to build search indexes and to fetch a page live when someone asks about it. Then they quote short passages from the pages they could read. A site can rank well in Google and still be invisible to all of this: a robots.txt rule that blocks one bot, a firewall that challenges another, content that only appears after scripts run, or structured data that does not parse.
Each tool below tests one of those failure points. You enter a URL or a domain, we fetch it the way a crawler would, and you get a list of checks marked pass, warning or fail. Every warning and failure comes with a specific fix, not a generic tip. The scoring is deterministic, so the same page gives the same result every time, and you can see exactly what was checked.
If you only have ten minutes, run these four in order. Together they answer the first questions that decide whether AI systems can use your site at all.
Confirm that the crawlers which build AI search indexes are allowed to fetch your important pages.
Check that your content is in the HTML before JavaScript runs, because most AI crawlers do not run scripts.
Make sure your structured data parses and tells machines who you are and what each page is about.
See which passages on a page are easy for an answer engine to quote, and which need rewriting.
Can AI crawlers reach and read your pages at all?
Five deterministic checks on one page: whether it is crawlable without JavaScript, whether its headings are in order, whether it carries structured data, whether an llms.txt file exists and whether the basic metadata is complete. You get a score out of 100 and a letter grade, and the same grade scale is used in the paid audit.
Open Free AI Visibility CheckRequests your page as GPTBot, ClaudeBot, PerplexityBot and other AI crawlers using each vendor's real user agent, then repeats the request as a normal browser. When a bot is blocked and the browser is not, the block usually comes from a CDN or firewall rule rather than from robots.txt, and the report says which it was.
Open AI Crawler Access CheckerShows the text a crawler recovers from your page without running JavaScript, next to the structure it finds: word count, heading outline and whether the page is an empty application shell. It is the quickest way to see what an AI system can actually quote.
Open Crawler ViewPaste a page URL and see which of 20+ crawlers your robots.txt allows to fetch it, and exactly which rule decided each answer.
How it works. We request /robots.txt from your site with our own honest user agent and parse every group and rule. Crawlers pick the group that names them, then the longest matching Allow or Disallow rule. If an Allow and a Disallow match equally, Allow wins. We follow that exactly. You see allowed or blocked for every crawler, the rule that decided it, and a fix when search or live-browsing bots are blocked.
Why it matters. AI search products use separate crawlers for training, search indexes and live browsing. Blocking a training crawler is a business choice. Blocking a search or live-browsing crawler removes you from the answers people are getting right now. A single stray Disallow can do that without anyone noticing.
Open Robots.txt Tester for AI CrawlersWe fetch your page the way an AI crawler does and measure what is already in the HTML before any script runs.
How it works. One request, no JavaScript, the same view GPTBot, ClaudeBot and PerplexityBot get. Word count, headings, links, an empty app container, and framework signatures such as Next.js, React, Vue or Angular. We look for text that sits in hydration data but not in the visible HTML. This is an estimate: we do not run a headless browser, and the report says so.
Why it matters. Most AI crawlers fetch HTML and stop. If your headline, pricing or product copy is added by JavaScript, those crawlers see an empty shell and have nothing to quote. Googlebot renders JavaScript, which is why this problem hides in normal SEO checks.
Open JavaScript Rendering CheckerCommon Crawl is the open web archive many language models learn from. Check whether your domain is in the latest crawls.
How it works. We check the three most recent crawls for your domain and its subdomains. Which crawls include you, a size signal and sample URLs. CCBot is Common Crawl's crawler. If you block it, future crawls skip you.
Why it matters. Many large language models were trained at least partly on data derived from Common Crawl. Being absent from it is one reason a model may know little about a brand. Being present is not a guarantee, and blocking CCBot is a legitimate choice.
Open Is My Site in AI Training Data?Is your content structured so answer engines can quote it?
Enter a URL and we extract every JSON-LD and microdata item, check required and recommended fields, and hand you a corrected JSON-LD block.
How it works. We extract JSON-LD blocks (including @graph) and microdata exactly as a crawler receives them, without running JavaScript. Organization, WebSite, Product, Article, FAQPage, BreadcrumbList, LocalBusiness and SoftwareApplication are checked for required and recommended properties. Missing properties are filled with clearly marked REPLACE placeholders so you can paste a complete block into your page.
Why it matters. Structured data tells machines who you are and what a page is about without guessing. Organization markup with sameAs links helps AI systems tie your site to the right entity. Broken JSON-LD is ignored silently, so errors cost you without any warning.
Open Schema Markup ValidatorSee your title, description, canonical, robots and social tags with character counts, and how the page looks as a Google result and a shared link.
How it works. Title, meta description, canonical link, robots meta, Open Graph and Twitter tags from the served HTML. Titles over about 60 characters and descriptions over about 160 get cut in results. Missing og:image is the usual reason shared links look bare. Google result, LinkedIn post, X card and WhatsApp link, built from the tags you actually have.
Why it matters. Titles and descriptions are the first thing both search engines and AI assistants read when summarising a page. Clear, accurate meta tags make your page easier to cite correctly, and good social cards decide whether a shared link gets clicked.
Open Meta Tags and Social PreviewWe split your page into passages by heading and score each one on traits that make text easy for an answer engine to lift and cite.
How it works. The same reader-mode extraction browsers use removes navigation and boilerplate. Each passage gets points for standing alone, concrete facts, naming the entity, question-style headings, a quotable length and lists or tables. You get the weakest passages first, with a specific rewrite suggestion for each problem.
Why it matters. Answer engines quote chunks, not pages. A paragraph that starts with "This also helps" cannot be quoted on its own. Passages with names, numbers and a clear answer are the ones that get lifted into responses.
Open AI Citability AnalyzerMake a 1200x630 social card from a page URL or your own headline. Pick a template and download the PNG.
How it works. We read the title, description and favicon from your page, or you type your own. Four flat designs: brand, dark, minimal and split. Edit the text and watch the card update. Upload it to your site and point og:image at it. We do not store your card.
Why it matters. A missing or generic share image makes links look untrustworthy on LinkedIn, Slack and WhatsApp, and chat products often show the same image when they link to your page.
Open OG Image GeneratorIs your brand a recognised entity? Look it up on Wikidata and Wikipedia and see whether your site links to it.
How it works. We look for items matching your brand name and prefer the one whose official website matches your domain. Official website, logo, founding date, industry and its English Wikipedia link. If you give a website, we check that your Organization JSON-LD points to the Wikidata item.
Why it matters. Knowledge graphs help AI systems tell you apart from similarly named companies. A Wikidata item and matching sameAs markup are strong, machine-readable proof of which entity your site represents.
Open Brand Entity CheckerSpeed, authority and presence in the data AI learns from.
Get mobile and desktop Lighthouse scores, Core Web Vitals and the top opportunities for any public URL, powered by Google PageSpeed Insights.
How it works. Two runs, mobile and desktop, using the official API. Performance, accessibility, best practices and SEO, plus LCP, CLS and TBT checked against Google's thresholds. If Google has enough traffic data for the URL we include the Chrome User Experience Report field results.
Why it matters. Slow pages are crawled less and abandoned more. Fast server response and light pages also make it cheaper for any crawler to read you fully.
Open PageSpeed CheckerHelp search and AI systems find what you publish.
Reads your sitemap, fetches the title and description of each page and writes a valid llms.txt grouped by section. It also compares the result with any llms.txt you already publish. The file is a proposed convention, not a standard, and the tool says so.
Open llms.txt GeneratorFollow a URL hop by hop. See every status code, how long each hop took, and the headers on the final response.
How it works. Each Location header is checked and requested separately, up to 10 hops, with timing for every step. Length, temporary versus permanent redirects, http to https, loops, slow hops and a failing destination. X-Robots-Tag, Strict-Transport-Security, Cache-Control, Content-Type and any canonical Link header.
Why it matters. Crawlers give up on long chains, and a hidden X-Robots-Tag: noindex can keep a page out of search and AI answers with no sign in the HTML. Redirect problems are invisible in a browser because it hides the hops.
Open Redirect Chain CheckerEnter a domain. We find your sitemaps, validate the XML, count URLs and check a sample for errors and hreflang mistakes.
How it works. From the Sitemap lines in robots.txt, then /sitemap.xml and /sitemap_index.xml. Sitemap indexes and gzip files are handled. Well-formed XML, URL counts against the 50,000 limit, absolute same-host URLs. Up to 50 URLs are requested for status codes, and hreflang alternates are checked for return tags.
Why it matters. A sitemap is the cleanest list of what you want found. Dead or redirecting URLs in it teach crawlers to trust it less, and broken hreflang sends the wrong language version to the wrong audience.
Open Sitemap ValidatorVerify your IndexNow key file, then tell participating search engines about up to 100 new or changed URLs at once.
How it works. Pick a key of 8 to 128 characters and host a text file containing only that key at your domain root. We can generate a key for you. We fetch the file and check that it holds exactly your key. Only after verification, we send up to 100 URLs from your domain to api.indexnow.org.
Why it matters. Bing powers parts of ChatGPT search and other assistants. IndexNow lets Bing and other participating engines learn about changes in minutes instead of waiting for a recrawl, so fresh pages get picked up sooner.
Open IndexNow SubmitterYes. Every tool runs without an account, a credit card or a trial. We limit how many checks one network can run in a short window so the tools stay available for everyone, and some results are cached for 24 hours.
No. Every score comes from fixed rules in code, so the same page gives the same result each time. The tools fetch your pages or call a public data source such as Google PageSpeed Insights, Common Crawl or Wikidata, and nothing is sent to an AI model.
We fetch the page you give us once or a few times to run the check, using our own user agent, GetBotRankBot, which is described on our bot page. Results are stored for 30 days so you can share a report link. We never fetch private or internal addresses.
It is optional. If you want a copy of the report and a weekly re-check that only emails you when something changes, enter your address after you see the result. You never need it to see the result itself.
Classic SEO tools are built around Googlebot. These are built around AI crawlers and answer engines, which behave differently: most do not run JavaScript, they use separate crawlers for training, search and live browsing, and they quote passages rather than ranking whole pages.
Yes. All 17 tools are also available through a free remote MCP server, so an assistant that supports the Model Context Protocol can run them directly. The same rate limits apply.
The free tools check access and readability. They cannot tell you whether ChatGPT, Claude or Perplexity actually mention your brand when people ask about your category. The GetBotRank visibility scan asks those engines real questions and tracks the answers over time. Run a full AI visibility scan.