AI Visibility Checker
Can AI read your website?
Enter your site's address to see whether ChatGPT, Claude, Gemini and Perplexity can get into your site and understand your content. We give you a score out of 100 and the fix for every problem.
What Does the AI Visibility Checker Look At?
Your score is the sum of seven categories. The checks that decide whether AI can get into your site carry the most points; the newest agent standards carry the fewest.
- 20 points
Crawl access
robots.txt can be read, AI crawlers aren't blocked, no firewall or bot protection turns them away, and you use Content-Signal to say how your content may be used.
- 18 points
Readability
The main text is in the HTML without needing JavaScript and is a reasonable length; the page has a title, a description, a canonical URL and a single H1.
- 16 points
Discoverability
The sitemap has lastmod dates and is declared in robots.txt; a missing page returns a real 404; the home page loads over HTTPS without a redirect chain.
- 14 points
Structured data
A complete Organization JSON-LD, your social profiles via sameAs and more than one schema type. LocalBusiness is only needed for local businesses and doesn't count towards the score.
- 12 points
Business information
Links to your about, contact and legal pages, machine-readable contact details and prices visible without signing in.
- 10 points
Agent-ready content
A well-formed llms.txt, a Markdown version of the page on request and Link headers that point to both.
- 10 points
Agent interaction
WebMCP tools for agents in the browser and a published MCP server card. Very few sites have these yet.
Which AI Crawlers Do We Check?
For each crawler, we check whether robots.txt allows it and whether your site actually responds to it. One company's crawlers do different jobs: blocking one doesn't block the other.
OAI-SearchBotOpenAI SearchChatGPT-UserOpenAI User requestGPTBotOpenAI Model trainingClaude-SearchBotAnthropic SearchClaude-UserAnthropic User requestClaudeBotAnthropic Model trainingPerplexityBotPerplexity SearchPerplexity-UserPerplexity User requestGoogle-ExtendedGoogle Model trainingApplebot-ExtendedApple Model trainingMeta-ExternalAgentMeta Model trainingCCBotCommon Crawl Model training
- Search
- Builds the index of pages the AI cites as sources in its answers. Block it and your link won't appear in answers.
- User request
- Opens your page on the spot when someone gives your address in a chat or when the AI looks something up for an answer.
- Model training
- Collects content to train new models. Blocking it doesn't affect the search and user-request crawlers.
How to Improve Your AI Score
Points are usually lost in one of the areas below. Start with the category where your result lost the most points, then scan your site again once it's fixed.
Allow AI crawlers
Remove the robots.txt lines that block OAI-SearchBot, ChatGPT-User, Claude-SearchBot or PerplexityBot, and check that your firewall or your CDN's bot protection isn't turning these crawlers away. If you don't want your content used for model training, keep search open and switch off training only: Content-Signal: search=yes, ai-train=no
Make your pages easy to find
Publish an XML sitemap, add a lastmod date to every URL and declare the sitemap in robots.txt with a Sitemap: line. Pages that don't exist should return a real 404, and your home page should load over HTTPS in a single step.
Make your content readable without JavaScript
Most AI crawlers don't run JavaScript; they read the first HTML the server sends. Put your main text in that HTML, and give every page a title, a description, a canonical URL and a single H1.
Say who you are
Add Organization JSON-LD to your home page: your name, logo, address, phone number and your social profiles via sameAs. This is how AI tells you apart from other businesses with the same name. If you have a place people visit, such as a shop, office or clinic, add the LocalBusiness type too; it's optional and doesn't count towards the score.
State your business details clearly
Link to your about, contact and legal pages from the home page, and write your phone number and email as text, not as an image. Make your prices visible without signing in and without needing JavaScript.
Add llms.txt and Markdown
Put an llms.txt that summarises your site in the root directory; when an agent asks for Markdown, serve the Markdown version of the page, and point to both with a Link header. These standards are new and optional; they give an agent a short, clean summary of your site.
What Is AI Visibility?
People looking for a product or service now ask ChatGPT, Gemini or Perplexity as well as a search engine. To answer, these tools read websites with their own crawlers; they can't cite a site they can't get into or can't understand.
AI visibility is how open and understandable your site is to these crawlers. Ranking high on Google doesn't guarantee it: a single line in robots.txt, bot protection or content that only loads through JavaScript can make your site invisible to AI.
- The user asksAsks ChatGPT for a recommendation
- The AI readsIts crawler visits your site
- It cites youPicks from the sites it could read
How to Read Your AI Score
72/100Score: The sum of the seven categories, out of 100. The checks that decide access carry the most points.BGrade: Your score as a letter: A for 90 and above, B for 75, C for 60, D for 40 and above; anything lower is F.17/20Category score: The points earned in a category against the most it can give. Start with the category where you lost the most points.GPTBot: AllowCrawler access: Whether robots.txt opens the home page to that crawler. If the site returns an error to that crawler, it counts as blocked even when it's allowed.llms.txt: 404Check: The result of each check: Passed, Attention, Problem or Not checked. Every check that doesn't pass shows its fix and a prompt you can copy underneath.
A check that can't apply —for example the format of llms.txt when there's no llms.txt— shows as Not checked. Run the check again after you change something on your site; the result shows the current state.
Visibility Check or Visibility Tracking?
They answer different questions. The check comes first: AI can't cite a site it can't read.
| Topic | Visibility check (this tool) | Visibility tracking |
|---|---|---|
| Question it answers | Can AI read my site? | Does my brand appear in AI answers? |
| What does it look at? | Your home page, robots.txt, sitemap and llms.txt | Many questions put to AI, and the answers |
| What do you get? | A score out of 100, a grade and the fix for every problem | Mention count, your share against competitors and the tone of the answers |
| When? | First, and after every change to your site | Once AI can read your site |
| Cost | Free | Usually a paid subscription |
Frequently Asked Questions
About AI crawlers, the score, robots.txt and llms.txt.
-
It's a free scan that measures whether tools like ChatGPT, Claude, Gemini and Perplexity can get into your site and understand your content. It visits the site the way an AI crawler does, runs more than 25 checks across seven categories and gives you a score out of 100 along with the fix for every problem.
-
No. The check measures whether your site is open and understandable to AI. To see how often you're mentioned in answers, you need a tracking tool that asks many questions on a regular basis. But being readable comes first: AI can't cite a site it can't read.
-
OpenAI's OAI-SearchBot, ChatGPT-User and GPTBot; Anthropic's Claude-SearchBot, Claude-User and ClaudeBot; Perplexity's PerplexityBot and Perplexity-User; plus Google-Extended, Applebot-Extended, Meta-ExternalAgent and CCBot. For each one, we look at the robots.txt rule and at how your site responds to that crawler.
-
75 and above (B and A) shows that your site is open and understandable to AI crawlers. If you're below 60, start with the category where you lost the most points: often a single line in robots.txt or bot protection is enough on its own to pull the score down.
-
If you want to appear as a source in AI answers, allow the search and user-request crawlers (OAI-SearchBot, ChatGPT-User, Claude-SearchBot, PerplexityBot). If you don't want your content used for model training, you can block only the training crawlers (GPTBot, ClaudeBot, Google-Extended, CCBot) or add
Content-Signal: ai-train=noto robots.txt. -
llms.txt is a text file placed in your site's root directory that summarises your site for AI in Markdown: who you are, your key pages and a short description of each. It isn't a mandatory standard; it's easy to add and helps agents understand your site correctly.
-
Google's crawler runs JavaScript and has known your site for a long time. Most AI crawlers, on the other hand, read the page's first HTML and can get stuck on robots.txt or bot protection. Checks like llms.txt and structured data also measure how well AI understands you, not your search ranking.
-
No. It reads your home page, robots.txt, sitemap and llms.txt, and the response to an address that doesn't exist. It only looks at public content; it doesn't sign in, submit forms or change anything on your site.
Ready to Raise Your Score?
Linux Hosting plans where you can edit robots.txt, llms.txt and your site files yourself start at $0.49 a month, with setup support included.