Skip to content
Growzo · AI crawlers

Thirteen crawlers decide whether models can read you.

Growzo ships a canonical registry of the 13 AI crawlers that matter, and for each one it states what blocking it actually costs you. Blocking GPTBot does not block ChatGPT Search citations, and blocking Google-Extended does not touch Google Search — they are separate tokens with separate consequences, and one over-broad Disallow can remove a site from every AI surface at once while rank tracking shows nothing at all. Growzo reads your robots.txt the way Google does, and it never claims a crawler obeys it.
Three robots.txt rules and their real consequences: blocking GPTBot removes you from OpenAI training but leaves ChatGPT Search citations intact; blocking Google-Extended removes Gemini grounding but leaves Google Search untouched; one over-broad Disallow closes every AI surface at once while rank tracking shows no change.
The same word — blocked — costs three completely different things depending on the token.
01 · Thirteen bots

The registry, and what each block costs.

A crawler list without consequences is trivia. The column that decides anything is the last one: what you give up by writing a Disallow. Most robots.txt files that block AI agents were copied from a blog post, and the person who pasted them has no idea which of these thirteen lines they just switched off.

CrawlerOperatorWhat it is forWhat blocking it costs you
GPTBotOpenAIModel trainingExcluded from OpenAI training data
OAI-SearchBotOpenAIChatGPT Search indexNo ChatGPT Search citations at all
ChatGPT-UserOpenAILive fetch for a userChatGPT cannot open your page on request
ClaudeBotAnthropicModel trainingExcluded from Claude training data
Claude-WebAnthropicUser-triggered fetchClaude cannot retrieve the page for a user
anthropic-aiAnthropicLegacy training tokenExcluded wherever the old name is honoured
PerplexityBotPerplexityCited answer indexGone from Perplexity's citation list
Google-ExtendedGoogleGemini and Vertex groundingNo Gemini grounding; Search unaffected
BytespiderByteDanceTraining and product crawlExcluded from ByteDance AI surfaces
meta-externalagentMetaMeta AI training and answersExcluded from Meta AI answers
CCBotCommon CrawlOpen corpus others reuseRemoved from the dataset many models reuse
cohere-aiCohereTraining and retrievalExcluded from Cohere-derived assistants
Applebot-ExtendedAppleApple Intelligence trainingNo Apple Intelligence use; Siri unaffected
Growzo keeps this registry as data rather than prose so it can be evaluated against your live file rather than read as advice. Three of the thirteen belong to OpenAI alone, and they do three different jobs — which is the single most common place a robots.txt decision goes wrong.
02 · Training is not search

The distinction that costs the most money.

“Should we block AI crawlers?” is not one question. It is at least two, and the two have opposite answers for most businesses. Training access is what a model learns from months or years before a user ever asks about you. Search and fetch access is what lets an assistant find you, quote you and link to you today.

GPTBot is not OAI-SearchBot

GPTBot crawls for model training. OAI-SearchBot crawls the index ChatGPT Search cites from. ChatGPT-User fetches a page live when a user’s prompt causes a browse. Blocking the first is a defensible position on training. Blocking the second removes you from ChatGPT’s citations entirely, which is almost never what the person writing the rule intended.

Google-Extended is not Googlebot

Google-Extended is a robots.txt opt-out token for Gemini grounding and Vertex AI generative features. It is not a crawler in its own right and it is not a ranking signal. Disallowing it removes you from Gemini grounding while leaving Google Search — AI Overviews included, since those are served from the ordinary Googlebot index — completely untouched.

Applebot-Extended is not Applebot

The same split again, from a different vendor: Applebot serves Siri and Spotlight search, while Applebot-Extended is the opt-out for training Apple Intelligence. Blocking the extended token leaves Apple’s search surfaces working. Blocking the base one does not.

One Disallow can close everything

A single over-broad rule can remove a site from every AI surface at once — and the symptom is invisible in ordinary rank tracking, because rankings do not move. The loss shows up only as citations that never arrive, which is the hardest kind of loss to notice and the easiest to leave running for a year.

03 · Longest match wins

Your robots.txt, read the way Google reads it.

Precedence is not first-rule-wins and it is not last-rule-wins. Google evaluates every rule whose path pattern matches the request and applies the longest one; when an Allow and a Disallow match at the same length, the Allow wins. That is why a file can look restrictive and be permissive, or the reverse, and why reading it by eye is unreliable.

Request pathRules that matchLongest matchResult
/blog/geo-guideDisallow: / + Allow: /blog//blog/ — 6 charactersAllowed
/app/settingsDisallow: / + Allow: /blog// — 1 characterBlocked
/blog/draft-7Allow: /blog/ + Disallow: /blog/draft/blog/draft — 11 charactersBlocked
/blog/Allow: /blog/ + Disallow: /blog/draft/blog/ — 6 charactersAllowed
/xAllow: /x + Disallow: /xequal length — Allow winsAllowed
The last row is the tie-break rule: at equal specificity, Allow beats Disallow.
robots.txt — evaluated per agent
GPTBotdisallowed by /
OAI-SearchBotallowed by /blog/
PerplexityBotallowed by /blog/
Google-Extendeddisallowed by /
Googlebotallowed — token not matched
Illustrative file. Growzo evaluates your real one, agent by agent.

The sentence Growzo will not soften

We never claim a crawler obeys anything. robots.txt is voluntary; this reports what the file INSTRUCTS. No tool can measure compliance from outside, and one that presents an instruction as an observed fact has quietly swapped a guess for a measurement.

04 · Original research

The AI Crawler Blocking Index.

Growzo publishes its own longitudinal research on what the web actually instructs AI crawlers to do, built from repeated crawler-policy snapshots and shipped with a downloadable data.csv so the numbers can be checked rather than believed. It lives at /index/ai-crawler-blocking inside the Growzo app. What makes it worth citing is not the sample size — it is the four invariants encoded into how the numbers may be reported.

Unreachable is not blocked

A site that timed out, moved or returned a 500 has told you nothing about its crawler policy. Counting it as blocking inflates every blocking rate ever published, and it is the most common flaw in this genre of research.

A zero denominator yields null

Zero observations divided into anything is not zero percent — it is not a number. The index returns null and renders it as unmeasured, rather than drawing a reassuring line along the bottom of a chart.

Nothing is interpolated

Gaps between snapshots stay gaps. A smooth curve drawn through missing observations is a picture of the smoothing function, not of the web.

One run is never a trend

A single snapshot is a measurement of one moment. The index refuses to describe direction until it has enough runs to have seen one, which is the discipline the word “trend” is supposed to carry.

05 · The delivery half

llms.txt, and the part nobody else ships.

Plenty of tools will generate an llms.txt for you. Growzo parses, validates and generates them — and then also serves them: live per-customer llms.txt and llms-full.txt endpoints that carry your whole published corpus as tag-free, attributable prose. Generating a file is a formatting exercise. Delivering the text is what actually puts readable content in front of a model that has already decided to look.

2
Live endpoints
llms.txt as an llmstxt.org index, and llms-full.txt as the whole corpus.
0
Tags in the prose
Markup stripped, so what a model reads is sentences rather than a DOM.
1
Source of truth
The endpoints serve your published CMS corpus — they cannot drift from it.
llms-full.txt — served live
Index — llms.txtllmstxt.org format
Corpus — llms-full.txttag-free prose
Attribution line per documentkept
Regenerated on publishautomatic
Serving

One honest caveat, because it belongs here: llms.txt is a proposal, not a ranking mechanism, and Google has said it does not use it. It is worth shipping because it is cheap and because some clients do read it — not because it is a lever. The GEO tier covers what is actually measurable.

06 · AI-readiness audit

Four sub-scores, over ten pages. Exactly ten.

The audit returns four sub-scores over a sample capped at ten pages— for a project audit and for the public audit alike. That ceiling is stated here rather than buried, because the gap between “we audited your site” and “we audited ten of its pages” is the whole difference between a finding and an impression. Pages are fetched through an SSRF-guarded crawler that reads robots.txt and sitemap.xml, and findings come back with a severity and a 0-100 score.

AI-readiness — 10-page sample
Crawler access9 of 10 reachable
Structured data70 / 100
Passage citability58 / 100
Attribution signals64 / 100
Illustrative values. The sample ceiling is not illustrative — it is ten pages.

A sample described as a sample

Ten pages will find a robots.txt mistake, a missing canonical pattern or a template with no structured data, because those are class defects that repeat. Ten pages will not find the one broken URL in a thousand, and Growzo does not pretend otherwise.

SSRF-guarded fetching

The crawler refuses to be pointed at internal addresses. An auditing tool that will fetch any URL you hand it is a request-forgery gadget wearing a friendly name, and the guard is the reason this one is safe to run against a domain you do not own.

Crawler access is the cheapest thing on this page to get right and the most expensive to get wrong, which is why it comes first in the audit. If you want the spend side of the same story, the hard money cap explains what every probe and fetch costs before it happens.

07 · Questions

Crawler questions, answered plainly.

Does blocking GPTBot stop ChatGPT from citing my site?

No, and this is the most expensive misunderstanding on the list. GPTBot is OpenAI's training crawler. OAI-SearchBot is the crawler behind the index ChatGPT Search cites from, and ChatGPT-User is the agent that fetches a page live when a user's prompt causes a browse. They are three separate tokens with three separate consequences: blocking GPTBot removes you from training data and leaves ChatGPT Search citations untouched, while blocking OAI-SearchBot removes you from those citations entirely.

What does blocking Google-Extended do to my Google rankings?

Nothing. Google-Extended is a robots.txt token that controls whether your content is used for Gemini app grounding and Vertex AI generative features. It is not Googlebot, it is not a ranking signal, and Google Search — including AI Overviews, which are served from the ordinary Search index via Googlebot — is unaffected by it. That asymmetry is exactly why the decision deserves to be made deliberately rather than inherited from a copied robots.txt file.

Does Growzo guarantee that a crawler obeys robots.txt?

No, and it will never say so. robots.txt is a voluntary protocol with no enforcement mechanism. Growzo reports what your file INSTRUCTS each agent to do, using the same longest-match precedence Google documents. Whether a given operator honours that instruction is not something any tool can measure from the outside, and a tool that reports 'blocked' as a fact is reporting its own assumption.

What is the AI Crawler Blocking Index?

Original longitudinal research: repeated snapshots of what real sites instruct AI crawlers to do, stored as crawler policy snapshots and published with a downloadable data.csv so the numbers can be checked rather than believed. It encodes four honesty invariants — an unreachable site is not a blocking site, a zero denominator yields null rather than zero percent, nothing between observations is interpolated, and a single run is never reported as a trend.

What does Growzo do with llms.txt?

Both halves. It parses, validates and generates llms.txt files, and it serves live per-customer llms.txt and llms-full.txt endpoints carrying your whole published corpus as tag-free, attributable prose. The generation half is common; the delivery half is not, and delivery is the half that actually puts readable text in front of a model.

How many pages does the AI-readiness audit check?

A maximum of ten, for a project audit and for the public audit alike. It returns four sub-scores over that sample. It is a sample by design and is described as one — it is not a site-wide crawler, and no page here will imply otherwise.

Find out what your file is really telling them.

Growzo runs on its own application and its own sign-in — an Autocloz account is not required. Point it at your robots.txt and read the thirteen answers, agent by agent.

Start freeTalk to us