AEO audit: what to check, and a free one you can run now

The verdict

1. The free audit, then the fixes done for you

   supereoour pick, and we make it

   • best for teams that want AEO done, not a dashboard: supereo does everything it takes to get you recommended, from your site and pages to the roundups, placements and videos AI reads;
   • it proves each move against control questions, with sample sizes and ranges on every number;
   • run it from Claude or ChatGPT: an MCP server with the same tools supereo's own agent uses.

   From $149 a month

   Run the free auditFree, no card, every answer shown.

The steps

1. Write the questions your buyers ask

   Start with what a buyer asks before they know your name: the best tools for their job, alternatives to what they use now, how two options compare, how to solve the problem you solve. Keep the questions that name you apart, since they can't show whether an engine finds you on its own: supereo reports them beside your rate, never in it^[2].

2. Ask each engine the way a buyer reaches it

   An app decides for itself when to search the web: Claude searches when it helps^[3], while through OpenAI's API the model searches only once web search is turned on, and even then chooses whether to^[4]. So an audit run only through APIs can miss what buyers see in the apps. supereo's free audit labels every number with how it asked^[1].

3. Ask each question more than once

   SparkToro found AIs don't give consistent lists of brand recommendations^[5]. Ask each question again on other days and read the pattern, never one answer. Check both of Google's surfaces too: Google says AI Mode and AI Overviews may use different models, so the responses and links they show will vary^[6].

4. Score every answer three ways

   An answer can name you without recommending you, or cite your page while it recommends someone else. Score each answer for named, recommended and cited, for you and each competitor, as supereo's judge does^[7]. Note when Google shows no AI Overview at all, which is often^[8].

5. List the pages the engines cite

   The pages an engine cites for your buyers' questions are the pages to beat or to get onto: your docs and pricing pages, lists and reviews on other sites, Reddit threads, YouTube videos. The AEO statistics guide counts them by type for every engine.

6. Check which AI crawlers may read your site

   Read your robots.txt for each search crawler: OAI-SearchBot for ChatGPT search^[9], Claude-SearchBot for Claude^[10], PerplexityBot for Perplexity^[11], and Googlebot for Google's AI features^[12]. Google-Extended decides whether Gemini may train on your pages and ground its answers in them^[13].

7. Check your CDN and firewall

   A block at your CDN or host stops a crawler your robots.txt allows: Google's own checklist for its AI features asks for crawling allowed in robots.txt and by any CDN or hosting infrastructure^[14]. The section on CDNs below has what to look for.

8. Read each page the way a crawler without JavaScript does

   Fetch each page as an AI crawler and read the HTML it gets back. Vercel found that none of the major AI crawlers render JavaScript^[15], so text that only appears after scripts run is text they never read.

9. Check each page against its question

   For each question, find the page that should answer it, then check that the answer is in its first lines, the question's words are in its title, it shows when it was updated, and its prices and comparisons are in plain text. The real run below checks one page this way.

10. Put the fixes in order, then prove them

    Fix what stops crawlers first, then the pages that should answer your most valuable questions, then your place on the pages engines cite. Then grade each fix against control questions it couldn't move, which carry everything the change had no part in^[16].

Real runs

{
  "url": "https://supereo.app/guides/can-claude-code-do-geo"
}

{
  "origin": "https://supereo.app",
  "path": "/guides/can-claude-code-do-geo",
  "robotsTxt": {
    "status": "ok",
    "detail": "robots.txt read."
  },
  "crawlers": [
    {
      "bot": "GPTBot",
      "aiCrawler": true,
      "allowed": true,
      "rule": "Allow: /",
      "group": "*"
    },
    {
      "bot": "OAI-SearchBot",
      "aiCrawler": true,
      "allowed": true,
      "rule": "Allow: /",
      "group": "OAI-SearchBot"
    },
    {
      "bot": "ChatGPT-User",
      "aiCrawler": true,
      "allowed": true,
      "rule": "Allow: /",
      "group": "*"
    },
    {
      "bot": "ClaudeBot",
      "aiCrawler": true,
      "allowed": true,
      "rule": "Allow: /",
      "group": "*"
    },
    {
      "bot": "Claude-SearchBot",
      "aiCrawler": true,
      "allowed": true,
      "rule": "Allow: /",
      "group": "Claude-SearchBot"
    },
    {
      "bot": "Claude-User",
      "aiCrawler": true,
      "allowed": true,
      "rule": "Allow: /",
      "group": "*"
    },
    {
      "bot": "PerplexityBot",
      "aiCrawler": true,
      "allowed": true,
      "rule": "Allow: /",
      "group": "PerplexityBot"
    },
    {
      "bot": "Google-Extended",
      "aiCrawler": true,
      "allowed": true,
      "rule": "Allow: /",
      "group": "*"
    },
    {
      "bot": "Googlebot",
      "aiCrawler": false,
      "allowed": true,
      "rule": "Allow: /",
      "group": "Googlebot"
    },
    {
      "bot": "Bingbot",
      "aiCrawler": false,
      "allowed": true,
      "rule": "Allow: /",
      "group": "Bingbot"
    }
  ],
  "llmsTxt": {
    "served": true,
    "status": 200,
    "chars": 3138
  },
  "edge": {
    "connected": false,
    "days": 7,
    "blockedFromOperatorAddresses": [],
    "spooferDenials": 0,
    "note": "Not connected yet. This needs supereo to read the site's own traffic at Cloudflare or Vercel: connect it at /dashboard/reads/connect in supereo."
  },
  "note": "From robots.txt and llms.txt alone. A block at the site's CDN or firewall, or an answer that depends on the crawler's address, shows only in its traffic."
}

{
  "url": "https://supereo.app/guides/can-claude-code-do-geo",
  "query": "can claude code do geo"
}

{
  "url": "https://supereo.app/guides/can-claude-code-do-geo",
  "finalUrl": "https://supereo.app/guides/can-claude-code-do-geo",
  "status": 200,
  "ms": 29,
  "signals": {
    "title": "Can Claude Code do GEO? A working setup (2026), with a real run",
    "titleYear": 2026,
    "h1": "Can Claude Code do GEO? Yes, with an MCP server",
    "h1Count": 1,
    "words": 2204,
    "jsShell": false,
    "h2": 9,
    "h3": 12,
    "questionHeadings": 7,
    "listItems": 29,
    "orderedListItems": 10,
    "tables": 1,
    "externalLinks": 1,
    "dollarFigures": 6,
    "percentFigures": 0,
    "bestForLabels": 2,
    "structuredData": [
      "Organization",
      "Person",
      "SoftwareApplication",
      "Offer",
      "UnitPriceSpecification",
      "Article",
      "BreadcrumbList",
      "ListItem",
      "FAQPage",
      "Question",
      "Answer"
    ],
    "author": null,
    "newestDate": {
      "date": "2026-11-04T00:00:00.000Z",
      "source": "json_ld",
      "ageDays": 3
    },
    "noindex": false,
    "canonical": "https://supereo.app/guides/can-claude-code-do-geo",
    "queryInTitle": 1,
    "queryInOpening": 0.6666666666666666,
    "openingText": "Yes, once it has the tools. Claude Code already reads and edits your code. Connected to supereo's MCP server it can also check which AI crawlers may read your site, fetch a page as ChatGPT's search cr",
    "aiCrawlersBlocked": [],
    "llmsTxtServed": true
  },
  "issues": []
}

{
  "url": "https://supereo.app/guides/can-claude-code-do-geo"
}

{
  "url": "https://supereo.app/guides/can-claude-code-do-geo",
  "finalUrl": "https://supereo.app/guides/can-claude-code-do-geo",
  "status": 200,
  "problem": null,
  "kind": {
    "value": "how_to",
    "source": "content_read"
  },
  "pageType": "docs",
  "score": 100,
  "checks": [
    {
      "id": "server_rendered",
      "label": "Real text in the HTML, not a JavaScript shell",
      "pass": true,
      "detail": "2204 words in the HTML"
    },
    {
      "id": "ordered_steps",
      "label": "Ordered steps",
      "pass": true,
      "detail": "10 ordered list items"
    },
    {
      "id": "answer_first",
      "label": "The answer in the first 200 characters under the H1",
      "pass": true,
      "detail": "answers up front: yes; opens \"Yes, once it has the tools. Claude Code already reads and edits your code. Connected to s…\""
    },
    {
      "id": "title_matches_question",
      "label": "A title matching the question",
      "pass": null,
      "detail": "needs the query"
    },
    {
      "id": "fresh",
      "label": "Updated within 30 days",
      "pass": true,
      "detail": "newest date 2026-11-04 (json_ld), 3 days old"
    },
    {
      "id": "specific",
      "label": "Specific rather than generic",
      "pass": true,
      "detail": "specificity 4 of 5, generic AI style 3 of 5"
    }
  ],
  "gaps": [],
  "numbers": {
    "dollarFigures": 6,
    "percentFigures": 0,
    "newestDate": {
      "date": "2026-11-04T00:00:00.000Z",
      "source": "json_ld",
      "ageDays": 3
    }
  },
  "read": {
    "status": "done",
    "summary": "A guide showing Claude Code can do GEO via supereo's MCP server, with a recorded run and setup steps; the publisher makes supereo and ranks it first.",
    "answersUpFront": "yes",
    "specificity": 4,
    "originalData": true,
    "sourcesCited": true
  },
  "costCents": 0.08085
}

What an AEO audit adds to an SEO audit

An SEO audit reviews how search engines crawl, index and rank your site. Google lists what such services cover, starting with a review of your site's content or structure^[17] and technical advice on hosting, redirects, error pages and the use of JavaScript^[18], and says an audit should give you realistic estimates of improvement and of the work involved^[19].

An AEO audit keeps those checks, since AI engines read your site through crawlers too, and adds what only the answers show: whether the engines name, recommend and cite you for your buyers' questions, which pages they cite instead, and whether what they say about you is true.

Do AI crawlers run JavaScript?

Mostly not. In Vercel's study of AI crawler traffic, none of the major AI crawlers rendered JavaScript^[15]. Gemini is the exception, because it uses Googlebot's infrastructure, which renders it fully^[20].

So a page built in the browser can read as empty to ChatGPT's and Claude's crawlers while it looks complete to you and to Google. An AEO audit fetches each page without running scripts and counts the words in its HTML, as the real run on this page does for one page.

Is your CDN blocking AI crawlers?

It may be, and the defaults have moved. Cloudflare first changed its default to block AI crawlers unless they pay creators for their content^[21]; its documentation now says new domains allow search, agents and training by default^[22]. A site set up in between can carry either setting, and nobody tells you which.

Two traps catch sites that think they allow AI search. Cloudflare applies a block on training to crawlers used for both search and training^[23], and Perplexity notes that Cloudflare's Bot Fight Mode can still challenge its bots when a firewall rule allows them^[24]. Read your CDN's AI bot settings beside your robots.txt. A fetch with a crawler's user agent shows what your server sends that name, but whether the real crawler gets through shows only in your CDN's logs or firewall events for requests from its published addresses.

Does an AEO audit need llms.txt or special markup?

Not for Google. Google says you don't need new machine-readable files, AI text files or markup to appear in AI Overviews and AI Mode, and that there's no special structured data to add^[25].

So an audit that marks you down for a missing llms.txt file is scoring something Google says its AI features don't need. An llms.txt file is cheap to serve, and supereo.app serves one, but it is no substitute for pages crawlers can read, answers in their first lines and facts that match everywhere you're described.

How many questions and runs before a result means anything

Which questions you ask moves a result more than chance does^[26], so a first audit needs questions written from your buyers' side and enough of them to cover your categories, comparisons and problems. supereo's free audit writes 20 questions your buyers ask and puts each one to all five engines it reads^[27][1], a first read of where you stand in about two minutes^[28].

A verdict needs repeats. SparkToro's advice for knowing an AI's recommendations is to ask over and over, usually at least 60 to 100 times, and average them out^[29]. supereo's tracker asks every question on six engines three times a week^[30], and its tests of a change leave Google's AI surfaces out because they vary too much from ask to ask^[31].

Live answers or training data: what a free audit asks

Free audits don't all ask the same thing. HubSpot describes its free grader as a one-time check of what ChatGPT, Perplexity and Gemini say about you based on their training data^[32]. That shows what a model remembers, not what it finds when your buyer asks and it searches the web.

supereo's free audit asks live: the consumer apps for ChatGPT and Gemini, Google's live results, and the API for Perplexity, labeled on every number^[1], and it shows every answer. It takes no card, and the result is yours^[33]. Free AI visibility checkers compares the free ones on what each checks.

Google-Extended isn't the switch for AI Overviews

Some audit checklists treat Google-Extended as an AI search crawler to allow. Google says Google-Extended doesn't affect a site's inclusion in Google Search^[34], and that robots.txt rules for Googlebot are what control how a site is crawled for Search, its AI features included^[12].

What Google-Extended does control is whether Google may use your content to train its Gemini models and to ground Gemini's answers^[13]. So blocking it keeps your pages out of Gemini's training and out of the grounding behind Gemini's answers, without taking you out of AI Overviews or AI Mode.

An AEO audit template you can copy

One sheet for the answers: a row per question and engine, with the date, whether the answer named you, recommended you and cited you, which competitors it recommended, and the pages it cited.

One sheet for the pages: a row per page, with the question it should answer, whether each AI crawler may read it, the words in its HTML without JavaScript, whether the answer is in its first lines, the day it was last updated, and the fix it needs.

Fill both again on the same questions after each batch of fixes. The free audit on this page fills the first sheet for you, and the real run fills the second for one page.

Have it done: supereo

Questions

How we compiled this

Where this page's words say how an AI engine finds, reads or cites pages, the sentence links the engine's own documentation, or another source, and the day it was read.

The runs are real calls to supereo's MCP server on its staging environment, made as a house account, recorded as they happened: the arguments as sent, each answer as the client received it, its JSON indented and trimmed only where it says so, the time each call took and when it ran.

We make supereo, so it is the one tool this page names as a pick.

How supereo measures its own results, with every number it runs on: How supereo measures.

supereo (supereo.app) is an AI agent that does everything it takes to get your brand recommended by ChatGPT, Gemini, Perplexity, Claude and Google, on your site and off it: site fixes, new pages, roundups, placements and video. It proves what worked against control questions. Made by superagnt.

Sources

1. ChatGPT and Gemini (the consumer apps), Google AI Mode and AI Overviews (live results), and Perplexity (the API, labeled on every number). supereo.app, read 12 November 2026.
2. How often AI recommends you counts the questions that don't name you; the ones that do are reported beside it. supereo.app, read 7 November 2026.
3. Note: If you have the new Claude experience, there's no web search toggle. Claude searches the web when it helps. support.claude.com, read 7 November 2026.
4. Using the Responses API, you can enable web search by configuring it in the tools array in an API request to generate content. Like any other tool, the model can choose to search the web or not based on the content of the input prompt. developers.openai.com, read 7 November 2026.
5. The bottom line is: AIs do not give consistent lists of brand or product recommendations. If you don’t like an answer, or your brand doesn’t show up where you want it to, just ask a few more times. sparktoro.com, read 7 November 2026.
6. AI Mode and AI Overviews may use different models and techniques, so the set of responses and links they show will vary. developers.google.com, read 7 November 2026.
7. Every answer is read by supereo's judge, which makes three calls about your brand, and the same three about each competitor: supereo.app, read 7 November 2026.
8. AI Overviews are only shown when our systems determine that it is additive to classic Search, and as such, often don't trigger. developers.google.com, read 7 November 2026.
9. OAI-SearchBot is for search. OAI-SearchBot is used to surface websites in search results in ChatGPT’s search features. Sites that are opted out of OAI-SearchBot will not be shown in ChatGPT search answers, though can still appear as navigational links. developers.openai.com, read 7 November 2026.
10. Disabling Claude-SearchBot on your site prevents our system from indexing your content for search optimization, which may reduce your site's visibility and accuracy in user search results. support.claude.com, read 7 November 2026.
11. PerplexityBot is designed to surface and link websites in search results on Perplexity. It is not used to crawl content for AI foundation models. docs.perplexity.ai, read 7 November 2026.
12. AI is built into Search and integral to how Search functions, which is why robots.txt directives for Googlebot is the control for site owners to manage access to how their sites are crawled for Search. developers.google.com, read 7 November 2026.
13. Google-Extended is a standalone product token that web publishers can use to manage whether content Google crawls from their sites may be used for training future generations of Gemini models that power Gemini Apps and Vertex AI API for Gemini and for grounding (providing content from the Google Search index to the model at prompt time to improve factuality and relevancy) in Gemini Apps and Grounding with Google Search on Vertex AI. developers.google.com, read 7 November 2026.
14. Ensuring that crawling is allowed in robots.txt, and by any CDN or hosting infrastructure developers.google.com, read 7 November 2026.
15. The results consistently show that none of the major AI crawlers currently render JavaScript. vercel.com, read 7 November 2026.
16. Difference in differences: how much the targets moved, minus how much the controls moved. The controls carry everything the change had no part in: a model update, a competitor's launch, a slow week. supereo.app, read 7 November 2026.
17. Many SEOs and other agencies and consultants provide useful services for website owners, including: Review of your site content or structure developers.google.com, read 7 November 2026.
18. Technical advice on website development: for example, hosting, redirects, error pages, use of JavaScript developers.google.com, read 7 November 2026.
19. An SEO audit should be about giving you realistic estimates of improvement, and an estimate of the work involved. If they guarantee you that their changes will give you first place in search results, find someone else. developers.google.com, read 7 November 2026.
20. Google's Gemini leverages Googlebot's infrastructure, enabling full JavaScript rendering. vercel.com, read 7 November 2026.
21. Cloudflare, along with a majority of the world's leading publishers and AI companies, is changing the default to block AI crawlers unless they pay creators for their content. blog.cloudflare.com, read 7 November 2026.
22. For new domains, Search, Agents, and Training use Allow (do not block) by default. developers.cloudflare.com, read 7 November 2026.
23. Training actions also apply to mixed-purpose crawlers used for both Search and Training. developers.cloudflare.com, read 7 November 2026.
24. Cloudflare’s Bot Fight Mode can’t be skipped with WAF custom rules. If Bot Fight Mode is enabled, it can still challenge Perplexity bots that match this rule. docs.perplexity.ai, read 7 November 2026.
25. You don't need to create new machine readable files, AI text files, or markup to appear in these features. There's also no special schema.org structured data that you need to add. developers.google.com, read 7 November 2026.
26. Which question is asked moves a number more than chance does, so the panel is versioned. supereo.app, read 7 November 2026.
27. 20 questions those buyers ask AI. You can edit them, and the competitors, after. supereo.app, read 7 November 2026.
28. About 2 minutes. The answers land on the page as they come in. supereo.app, read 7 November 2026.
29. My takeaway on trustworthiness is that if you really want to know an AI’s set of recommendations, you need to ask over and over again; usually at least 60-100X, then average these out sparktoro.com, read 7 November 2026.
30. Every question runs on six engines, 3 times a week each, and never twice on one engine within 30 minutes. supereo.app, read 7 November 2026.
31. One engine: ChatGPT unless you choose Gemini, Perplexity or Claude. Never Google's AI surfaces, which vary too much from ask to ask, or show no overview at all. supereo.app, read 7 November 2026.
32. HubSpot's AI Search Grader is a free, one-time check that reveals what ChatGPT, Perplexity, and Gemini say about you based on their training data. hubspot.com, read 7 November 2026.
33. Yes. The audit takes no card, and the result is yours. supereo.app, read 7 November 2026.
34. Google-Extended does not impact a site's inclusion in Google Search nor is it used as a ranking signal in Google Search. developers.google.com, read 7 November 2026.
35. In the EU, economic operators, who meet a range of criteria can obtain AEO status and in return they are entitled to certain benefits. taxation-customs.ec.europa.eu, read 7 November 2026.
36. an official examination of the quality or condition of something dictionary.cambridge.org, read 7 November 2026.
37. ChatGPT can search the web to answer questions with current information and links to relevant sources. It can also use location information to find local results, including restaurants, news, and weather. help.openai.com, read 7 November 2026.
38. Connect custom MCP servers to ChatGPT to use their tools in conversations. help.openai.com, read 7 November 2026.
39. Before making significant changes to your site based on a third-party tool's audit, be sure to check their recommendations against official guidance from Google Search, think critically about any claims or recommendations you hear, and make your own informed decisions. developers.google.com, read 7 November 2026.