Check AI visibility: Can ChatGPT read your website?

Are AI crawlers allowed in, do they see your text, do they understand who you are? A score from 0 to 100 in 30 seconds.

GPTBot
ClaudeBot
PerplexityBot
Gemini
Copilot
Google AI

Featured on

Related services for AI visibility

What AI crawlers see on your website

Six areas decide whether AI systems find, read and cite you. The check tests all of them – with read-only requests.

Bot access & robots.txt
We evaluate 16 crawlers from AI and search providers in your robots.txt, fetch your homepage live as GPTBot, ClaudeBot and PerplexityBot – and detect whether a bot blocker like Cloudflare quietly shuts them out.
llms.txt
The check looks for an llms.txt – a proposed convention (llmstxt.org), not an official standard, that tells AI systems which content to use and prioritize.
Sitemap & freshness
We read your XML sitemap, count its entries and inspect lastmod dates – the freshness signals AI crawlers use to decide what is worth indexing.
Structured data
JSON-LD schema, sameAs links and speakable markup make your entity unambiguous to AI. The check lists existing types and missing fields.
Server rendering
Content that only appears after JavaScript runs stays invisible to many AI crawlers. We measure how much text already sits in the delivered HTML (SSR).
Citability
Clear definitions, short paragraphs, lists and a clean heading hierarchy decide whether AI quotes your passages verbatim. The check scores exactly that.

Your result at a glance

Right after the check you see where your website stands and what to tackle first:

Overall score from 0 to 100, rated from “Critical” to “AI-ready”
Scores for all 9 categories – from bot access to citability
Every finding with raw data, such as the HTTP status GPTBot gets from your server
The biggest levers first, with the possible points gained per fix

Check AI visibility

Sample resultexample-shop.com

58/100Needs work

Who the check is for

Your question isn't here?

Ask directly

Online shops

Typical AI question: “Which shop has good e-bikes under €3,000?”
Typical blocker: Product data only via JavaScript, bot protection in the CDN

B2B & industry

Typical AI question: “Manufacturers of stainless steel tanks in NRW?”
Typical blocker: No schema, thin homepage, details only in PDFs

Service providers & agencies

Typical AI question: “Good tax advisor in Düsseldorf?”
Typical blocker: No LocalBusiness data, profiles not linked via sameAs

Why AI overlooks your website

The six most common blockers and how to fix them

GPTBot or ClaudeBot are blocked in robots.txt.

Explicitly allow the relevant AI providers' crawlers – otherwise you appear in no AI answer.

Content only shows up after JavaScript has loaded.

Server-side rendering (SSR) delivers the text straight in the HTML – readable for every bot, even without running JS.

There is no llms.txt and no schema markup.

Provide an llms.txt and add JSON-LD for organization, product and FAQ – so AI systems understand your entity.

Copy is promotional rather than factual and hard to quote.

Phrase answers in short, self-contained paragraphs with numbers, definitions and lists – the format AI models prefer to cite.

A Cloudflare or WAF rule blocks unknown user agents.

Allowlist AI bot user agents instead of rejecting them wholesale as scrapers.

High server response time (TTFB) makes crawlers give up.

Caching, a CDN and lean HTML push TTFB under one second – crawlers fetch more pages in less time.

Become AI-visible in 4 steps

Order matters: open the gate first, then structure content, then monitor.

Open bot access01

Allow GPTBot, ClaudeBot and PerplexityBot in robots.txt. Check Cloudflare, WAF and plugins for user-agent filters – common pitfall: Cloudflare defaults tend to reset on updates, so re-check after changes.

Build Schema.org and entity linking02

Add Organization, WebSite and BreadcrumbList schema. Use sameAs links to LinkedIn, Wikipedia and Crunchbase so AI can uniquely identify your brand.

Publish llms.txt03

Create a Markdown file at /llms.txt with a curated list of your most important URLs. Format per llmstxt.org – a proposed convention, not an official standard; some AI crawlers and research agents already read it.

Establish server rendering and monitoring04

Make sure main content lives in the initial HTML. Set up weekly monitoring that tracks bot responses, schema coverage and response times.

You have your score. Here is what comes next.

Every finding and fix is in the report. How you implement them is up to you.

Do it yourself
For teams with in-house development
The GEO guide bundles the key fixes from the check into an actionable playbook – free as a PDF.
Get the guide
Full audit
For shops and sites with many page types
We check every template, not just the homepage – category, product, blog, landing pages – and deliver a prioritized action plan.
Request an audit
Implementation & monitoring
For companies without in-house GEO capacity
Our GEO agency implements technology and content and keeps checking whether AI crawlers can still read your pages.
To the GEO agency

Check one part on its own: Check your llms.txt · May GPTBot & ClaudeBot crawl your site?

Technical check or mention tracking?

Two tools, two questions. Get the technical basis right first, then track – otherwise you measure symptoms instead of causes.

Can AI read and cite my website?AI Visibility Check

Tests bot access, delivered text, structured data and citability – one-off, instant and free.

Does AI actually mention my brand?Mention tracking

Tools like Peec AI, Profound or the Semrush AI Toolkit send ChatGPT & co. the same prompts on a schedule and measure how often your brand comes up.

GEO check vs. SEO audit: what is the difference?

A GEO check tests whether AI systems can read and cite your website as a source. An SEO audit tests how well you rank in Google. The levers overlap, but signals, bots and success metrics are their own thing.

Aspect	SEO audit	GEO check
Goal	Top-10 ranking on Google	Citation in ChatGPT, Claude, Perplexity
Key bots	Googlebot, Bingbot	GPTBot, ClaudeBot, PerplexityBot
Index file	sitemap.xml	llms.txt + sitemap.xml
Structured data	Rich snippets (Product, FAQ)	Entity linking (Organization, sameAs, @graph)
Rendering	Google renders JS — with delay	AI bots often skip JS — SSR mandatory
Success tracking	Search Console, rank trackers	Citation tracking, brand mentions in LLMs

How your AI visibility is measured

Your question isn't here?

Ask directly

Categories and weights

Ask ChatGPT, Claude or Perplexity for a provider and you get sources back. Whether your website can be one of them depends on technology first: are AI crawlers allowed in, do they see your text, do they understand who you are? The check answers that in 30 seconds – with a score from 0 to 100, every finding with its raw data and concrete fixes.

The score is built from 9 categories with weighted individual checks. A warning counts 40% of a check's points. Every check shows its raw data – you can trace and verify each finding yourself.

How your AI visibility is measured
Category	Weight
AI citability	22 %
Authority & trust	18 %
robots.txt	14 %
Bot blocker	12 %
Technical & performance	12 %
Homepage	10 %
Schema & structured data	9 %
Sitemap	2 %
llms.txt	1 %

Fundamental blockers cap the score: if all three simulated AI bots are blocked, it tops out at 39. A full robots.txt block caps it at 25, missing server rendering at 49.

What your score means

0–39CriticalAI crawlers can't reach your content or find barely any usable text.
40–59Needs workTypical non-optimized websites land here – the basics exist, but key AI signals are missing.
60–79Solid baseThe key requirements are met; some levers remain open.
80–100AI-readyTechnically well set up – AI systems can read and cite your content.

What the check does not measure

No actual mentions or citations in AI answers – that requires mention tracking. The check shows the technical blockers behind them.
Only the homepage plus robots.txt, llms.txt and sitemap.xml – no subpages.
Bot simulation via user agent from our IP, not IP-verified crawler identity.
No domain authority: a very well-known website would get cited despite a mediocre score.

What the check tests in detail

Each category bundles several individual checks. They all run on every scan – weighted by the share shown above.

AI citability: Fact density (numbers, percentages, years), definition and answer patterns, paragraph and passage length, lists and tables, clean heading hierarchy, question-style headings.
Authority & trust: Freshness and date signals, sameAs profile links, author/byline, contact details, outbound source links, about-us signals, experience signals, visible ratings, knowsAbout topics.
robots.txt: robots.txt reachability, access rules per AI crawler (GPTBot, ClaudeBot, PerplexityBot and more), training/scraper bots, wildcard blocks, sitemap reference, Content-Signal directives, noai meta tags.
Bot blocker: Live fetches as GPTBot, ClaudeBot and PerplexityBot – is the page served in full, suspiciously throttled, or blocked?
Technical & performance: Server rendering (content visible without JavaScript), security headers, canonical tag, viewport, compression, cache-control, image dimensions (CLS), alt text.
Homepage: HTTP status, response time (TTFB), HTTPS, X-Robots-Tag, amount of text, meta description, title tag, exactly one H1.
Schema & structured data: JSON-LD present and parses cleanly, recommended @types (Organization, WebSite, LocalBusiness, BreadcrumbList, FAQPage), required fields per type, invalid properties (markup errors), required fields for app rich results, speakable, Open Graph tags, hreflang.
Sitemap: sitemap.xml reachability, lastmod freshness signal, number of listed URLs.
llms.txt: Presence and formal validity of an llms.txt.

You can verify every finding yourself, for example the bot-blocker check in your terminal:

curl -A "GPTBot" -I https://your-domain.com

Version history of the test suite

The check is extended continuously. Recently added tests:

2026-07-01Schema markup errors: new checks for invalid properties (which Google Search Console reports as “markup errors”) and for missing required fields for app rich results (SoftwareApplication needs offers plus aggregateRating or review).
2026-06-11Weighted scoring model with published category weights, score caps as kill criteria, and the new “Authority & Trust” category (E-E-A-T signals).
2026-05-26Added the “Technical & Performance” and “AI Citability” categories with their individual checks.
Baserobots.txt evaluation, live bot access, homepage basics, core schema data, sitemap and llms.txt.

The key AI crawlers

The big providers run training and search through separate user agents. The check evaluates 16 crawlers in your robots.txt and also fetches your homepage live as GPTBot, ClaudeBot and PerplexityBot.

Frequently asked questions about the AI Visibility Check

Your question isn't here?

Ask directly

What is an AI Visibility Check?

An AI Visibility Check tests whether AI crawlers can fetch and read a website and how well its copy works as a source. Our check simulates the crawlers GPTBot, ClaudeBot and PerplexityBot, reads robots.txt, llms.txt and the sitemap, and rates the delivered HTML across 9 categories – from bot access and structured data to the citability of your copy. The result is a 0–100 score with raw data for every finding. Whether an AI actually cites your site is not something the check measures. Free, no email, in 30 seconds.

What is a GEO check?

A GEO check (Generative Engine Optimization) tests whether a website meets the technical requirements to appear as a source in answers from AI search systems such as ChatGPT, Perplexity or Google AI Overviews: bot access, server-rendered text, structured data and citable content. Our AI Visibility Check is a free GEO check for the homepage.

How do I check whether ChatGPT knows my website?

In three steps. First the technical basis: the AI Visibility Check shows whether ChatGPT & Co. can read and cite your website at all. Second a manual sample: ask ChatGPT, Perplexity and Gemini 10 to 20 real customer questions (“Which providers for X in Y do you recommend?”) and note whether you are mentioned or linked. Third mention tracking: tools such as Peec AI, Profound or the Semrush AI Toolkit repeat such prompts regularly and measure your share of voice against competitors.

How can I measure my AI visibility?

On two levels. Technically, the AI Visibility Check tests whether AI crawlers can read and understand your website. On the content side, mention-tracking tools show how often ChatGPT, Perplexity or Google AI Overviews actually name your brand. Only both together give the full picture.

What is LLM visibility and how does it differ from SEO?

LLM visibility describes whether and how your content appears in answers from ChatGPT, Claude, Perplexity, Gemini and Google AI Overviews. Classic SEO optimizes for Google rankings; LLM optimization optimizes for AI citations. The levers overlap (clean HTML, schema, speed) but the bots, the signals (llms.txt, entity linking, FAQPage schema) and the success metrics are their own thing.

What is the difference between a visibility index and AI visibility?

The visibility index is a metric from SEO tools such as Sistrix for a domain's visibility in classic Google rankings – calculated from the positions and search volume of a fixed keyword set. AI visibility, by contrast, describes whether a website appears as a source in answers from ChatGPT, Claude, Perplexity or Google AI. The two are related but measure different things. For a visibility index check you need an SEO tool; whether your website is readable and citable for AI is what the AI Visibility Check tests.

Why is LLM visibility especially important for online shops?

More and more shoppers ask ChatGPT or Perplexity instead of Google for product research. If you're not quoted or linked in those answers, you lose qualified traffic — even with strong Google rankings. Shops need server-rendered product data, Product and Offer schema, and they have to make sure AI bots aren't blocked at the WAF.

Which AI crawlers should I know about?

The most important: GPTBot (OpenAI training), OAI-SearchBot and ChatGPT-User (ChatGPT search and browsing), ClaudeBot, Claude-User and Claude-SearchBot (Anthropic), PerplexityBot and Perplexity-User (Perplexity), Applebot-Extended (Apple Intelligence), CCBot (Common Crawl, feeds many LLMs) and Bytespider (ByteDance/Doubao). Google-Extended is not a crawler of its own but an opt-out directive for Gemini training in robots.txt. Our check evaluates these and more – 16 crawlers in total – in robots.txt and fetches the homepage live as GPTBot, ClaudeBot and PerplexityBot.

Is it harmful to allow AI crawlers?

Not necessarily, because training and search can be controlled separately. At OpenAI, GPTBot collects training data while OAI-SearchBot feeds ChatGPT search; at Anthropic, it is ClaudeBot and Claude-SearchBot. If you want to stay out of training data but be cited in AI answers, block only the training crawlers.

What is llms.txt and do I need it?

llms.txt is a proposed convention (llmstxt.org): a Markdown file at /llms.txt that gives AI crawlers a curated list of your most important URLs – similar to sitemap.xml for Google. It's not an official standard, but some AI crawlers and research agents already read it. Minimal effort, clear upside.

Why do AI bots sometimes get different content than browsers?

Cloudflare Bot Fight Mode, ModSecurity, fail2ban, Varnish rules, Shopware plugins and nginx user-agent filters often block AI bots wholesale. Some setups respond with a captcha page, an empty fallback HTML or status 403. Browsers never see this — AI bots do. That's exactly the gap our bot-blocker check exposes.

Which Schema.org types matter most for AI?

Mandatory: Organization (who you are), WebSite (sitelinks search) and — if local — LocalBusiness or ProfessionalService. Strongly recommended: BreadcrumbList (hierarchy), FAQPage (voice/AI answers), Product + Offer (shops), Article + Author (blog/magazine), HowTo (guides). Combine in one @graph and enrich with sameAs links to LinkedIn, Wikipedia, Crunchbase, GitHub.

How often should I check my AI visibility?

At least after every release that touches routing, headers, robots.txt, schema or performance – and as a routine roughly once a month: CDN bot protection, Cloudflare defaults or shop plugins can change without you noticing. The AI Visibility Check keeps no history: note down the score or set up monitoring if you want to track the trend.

What is a good score?

From 60 points the key requirements are met; from 80 a website counts as AI-ready. Below 40, AI crawlers usually cannot reach the content at all or find little usable text. Typical unoptimised websites land between 40 and 59 points.

How quickly do improvements take effect?

Technical fixes such as an opened-up robots.txt take effect the next time the crawlers fetch your site. Whether and when an AI cites you afterwards also depends on content, authority and the model in question. In the check you see the effect immediately: every run reads the current state of your website.

Can nextlevels help me improve and monitor my LLM visibility?

Yes. We audit your full site (not just the homepage), build the missing JSON-LD, llms.txt and FAQ structures, clear bot blockers at the CDN and WAF layer, optimize server rendering and set up ongoing monitoring that tracks changes in robots.txt, bot responses, schema coverage and response times. Drop us a line about what you want to achieve — we'll get back to you within one business day.

Is the check worth it for B2B and industrial companies without an online shop?

Yes. Buyers and specialists increasingly research in ChatGPT or Perplexity – with queries like "manufacturer for X" or "provider for Y in NRW". Whether your website can appear as a source there depends on the same technical requirements as for shops: crawler access, structured data, citable content. Entity visibility matters regardless of whether you sell online.

Does the AI Visibility Check cost anything?

No. The AI Visibility Check is completely free – no email address, no sign-up, no automatic follow-up. We only store the checked domain, the score and the timestamp to evaluate the tool. Costs only arise if you commission us for a full audit or for implementing the fixes – that is your decision after the check.

Can I check a subpage as well?

The free check always analyses the homepage of the domain you enter, plus robots.txt, llms.txt and the sitemap. Product, category and blog pages have their own templates and therefore their own failure points – we check those in the full audit. Please only check domains you are responsible for.

What happens to my URL and the result?

The check only fetches publicly accessible files (robots.txt, llms.txt, sitemap.xml, homepage). We store the checked domain, score and timestamp to evaluate and improve the tool. No email address is required and there is no automatic follow-up contact. Details are in our privacy policy.