Free GEO Checker: Can AI Assistants Cite Your Page?
Free GEO checker. See whether ChatGPT, Claude, Perplexity and Google can fetch and cite your page, based on robots.txt, page content, structured data and llms.txt.
What is GEO (generative engine optimization)?
Generative engine optimization (GEO) is the work of making your content easy for AI assistants and AI search to find, read and quote in their answers. Classic SEO aims for a place in a list of links. GEO aims for your page to be one of the sources an assistant uses when it writes an answer, so the page has to be reachable by the bots these products run, readable without JavaScript and clear enough to quote.
This checker fetches the page, its robots.txt and its llms.txt. It then tests the page against 12 AI crawlers and runs 10 readiness checks. The checks are weighted to a total of 100 points. A pass earns full points, a warning half and a failure none, so the score is the share of points you earned. You get a score, a table showing whether each bot is allowed, and a note on what to change.
The checker reads only what a server sends. It does not run JavaScript, so its view matches what most AI crawlers see when they fetch the page.
Search bots and training bots
Each AI company runs more than one crawler, and the purposes differ. Search and answer bots fetch pages so an assistant can show or cite them in a live answer. Training bots collect content for building future models. Blocking a training bot does not stop an assistant from citing you, and blocking a search bot does. This tool tracks 7 search or answer bots and 5 training or control tokens, and it only lowers the score when a search bot is blocked.
The table below lists every bot the checker tests. Each row names the company, what the bot is used for, and whether it counts as a search bot.
| Bot | Owner | Used for | Type |
|---|---|---|---|
| OAI-SearchBot | OpenAI | ChatGPT search results | Search |
| ChatGPT-User | OpenAI | Fetches pages a ChatGPT user asks about | Search |
| GPTBot | OpenAI | Model training | Training or control |
| Claude-SearchBot | Anthropic | Claude search results | Search |
| Claude-User | Anthropic | Fetches pages a Claude user asks about | Search |
| ClaudeBot | Anthropic | Model training | Training or control |
| PerplexityBot | Perplexity | Perplexity search index | Search |
| Perplexity-User | Perplexity | Fetches pages a Perplexity user asks about | Search |
| Googlebot | Google Search and AI Overviews | Search | |
| Google-Extended | Gemini training and grounding (control token) | Training or control | |
| Applebot-Extended | Apple | Apple AI training (control token) | Training or control |
| CCBot | Common Crawl | Open dataset used by many models | Training or control |
The 10 readiness checks
The bot access check carries the most weight, because a blocked search bot can keep a page out of AI answers entirely. The other checks measure how easily a model can understand and quote the page.
| Check | Why it matters | Points |
|---|---|---|
| AI search bots allowed in robots.txt | If OAI-SearchBot, Claude-SearchBot, PerplexityBot or Googlebot are blocked, those assistants cannot cite the page. | 20 |
| Snippets allowed | nosnippet or max-snippet:0 stops search engines (and AI Overviews) from quoting the page. | 10 |
| Content is in the HTML | Most AI crawlers do not run JavaScript. Text added only by scripts is invisible to them. | 15 |
| Structured data | JSON-LD (Organization, Product, Article, FAQPage) states facts in a form machines read reliably. | 12 |
| Question-and-answer content | Answer engines favour pages that phrase headings as questions and answer them directly. | 10 |
| Clear description | A meta description and an early summary paragraph give AI a quotable one-line answer. | 8 |
| Dates are stated | Published and modified dates help engines judge whether information is current. | 7 |
| Author or organisation named | A named author or publisher is a trust signal for citation. | 6 |
| Specific facts and numbers | Pages with concrete numbers, lists and tables are easier to quote than general prose. | 6 |
| llms.txt file | A proposed standard (llmstxt.org) that gives AI assistants a markdown guide to the site. Optional, low cost. | 6 |
What your GEO score means
- 80 to 100Ready
- AI search and answer bots can reach the page, and the content is readable without JavaScript. Keep the page current and watch for new blocking rules.
- 50 to 79Partly ready
- The page can be read, but something limits it: a blocked search bot, text that needs JavaScript, or missing structured data and dates. Fix the highest-weight items first.
- 0 to 49Hard to cite
- AI assistants are likely to skip or misread the page. Usually the cause is a block on search bots, noindex or nosnippet, or almost no text in the HTML.
A high score does not guarantee a citation. It means the page is set up so that AI systems can reach and read it. Whether an assistant picks the page also depends on how well it answers the question and what other sources exist.
How to make a page easier for AI to cite
- Allow the search bots. Check
robots.txtfor OAI-SearchBot, Claude-SearchBot, PerplexityBot and Googlebot. Blocking Googlebot also stops Google from crawling the page, so it cannot appear in Google Search or AI Overviews. - Put the text in the HTML. Server-render the main content, so the answer is in the first response and not added later by a script.
- Answer questions directly. Use question headings, then give the answer in the first sentence or two under each one. A short summary near the top gives an assistant a quotable line.
- Add facts and dates. Include specific numbers, a comparison table or a list, and machine-readable dates in JSON-LD (datePublished, dateModified) or in a
<time datetime>tag. - Name the author or publisher. Add an Organization or Person object in JSON-LD and a visible byline on articles.
- Add an llms.txt file. Serve a markdown file at
/llms.txtwith a title and short links to your key pages. It is optional and not every engine reads it yet.
The page also needs to be indexed for search. Run the on-page SEO checker to check titles, headings and structured data, and use the Domain Rating checker to see how strong the linking domain is.
How the GEO / AI Citation Checker works
- We fetch the page, its robots.txt and its llms.txt.
- We test the page against 12 AI crawlers, including OpenAI, Anthropic, Perplexity, Google, Apple and Common Crawl, and run 10 readiness checks.
- You get a 0 to 100 score, a per-bot allow or block table and what to fix.
Frequently asked questions
Question 1: Does an allowed bot mean I will be cited?
No. Allowing a bot only gives it permission. Whether it fetches your page and cites it also depends on your server or CDN letting it through, and on how good the content is.
Question 2: Will blocking GPTBot hurt my search rankings?
No. GPTBot is a training crawler, and blocking it does not affect Google Search or ChatGPT search. Googlebot and OAI-SearchBot are the bots that matter for search and citations.
Question 3: Does a high score mean ChatGPT or Claude will cite my page?
No. The score shows whether AI systems can reach and read the page. It does not check whether any assistant has fetched or quoted it.
Question 4: Is llms.txt required?
No. It is an optional, proposed standard. The check is worth passing because it costs little, but no page loses citations for lacking one.
Question 5: Is it free?
Yes. It is free and needs no signup, with a limit of 6 checks per minute per visitor.