AI Crawler & Bot Directory

AI crawlers are automated agents that request public web pages for model training, answer-engine search indexes, structured datasets, or live assistant browsing. They use distinct User-agent names in robots.txt—different from classic search crawlers like Googlebot—and each major AI company may run several bots for different purposes. This directory compares them side by side so you can decide what to allow or block with clear, verifiable facts.

By Vinespire Editorial Team, Editorial · Reviewed by Vinespire Editorial Team, Editorial ·

Update: Registry verified against operator docs. Hub stamp 2026-07-19; each bot page carries its own lastVerified date.

See our sourcing methodology →

All AI bots

Filter by category or sort by operator. Every row is present in the initial HTML for crawlers; filters only hide rows in the browser.

Showing 29 of 29

Bot nameOperatorPurposeRespects robots.txt?Category
AI2Bot
AI2Bot
Allen Institute for AI (Ai2)Explores certain domains for web content used to train open language models at the Allen Institute for AIPartialTraining
Amazonbot
Amazonbot
AmazonAmazon’s crawler for improving products such as Alexa and AI experiencesYesTraining
Claude-SearchBot
Claude-SearchBot
AnthropicIndexes content to improve Claude search result relevanceYesSearch indexing
Claude-User
Claude-User
AnthropicFetches pages when people use Claude and need live web contentYesLive browsing
ClaudeBot
ClaudeBot
AnthropicCollects web content that may contribute to Claude model training and safetyYesTraining
Applebot
Applebot
AppleApple’s web crawler for Spotlight, Siri, Safari search features; crawled data may also support generative features unless opted out via Applebot-ExtendedYesSearch indexing
Applebot-Extended
Applebot-Extended
ApplePreferences for Apple Intelligence / generative features (related to Applebot)YesDataset / other
Bytespider
Bytespider
ByteDanceByteDance web crawler associated with AI / content productsYesTraining
cohere-ai
cohere-ai
CohereCohere crawler / bot identifier used in robots.txt for AI data preferencesYesTraining
CCBot
CCBot
Common CrawlBuilds the Common Crawl open web dataset used by many AI researchers and companiesYesDataset / other
Diffbot
Diffbot
DiffbotExtracts structured data from the web for knowledge graphs and AI productsYesDataset / other
DuckAssistBot
DuckAssistBot
DuckDuckGoReal-time crawler for DuckDuckGo AI-assisted answers that cite sources (not used to train AI models per DuckDuckGo)YesSearch indexing
DuckDuckBot
DuckDuckBot
DuckDuckGoPrimary DuckDuckGo search crawler for improving organic search results and the private search experienceYesSearch indexing
Google-CloudVertexBot
Google-CloudVertexBot
GoogleCrawls sites when site owners request website data for building Vertex AI Agents; does not affect Google Search rankingsYesDataset / other
Google-Extended
Google-Extended
GoogleControls use of content for Gemini and Google AI features (not classic Googlebot search)YesDataset / other
Googlebot
Googlebot
GooglePrimary Google crawler family for Google Search indexes and related Search surfaces (not Google-Extended generative preferences)YesSearch indexing
GoogleOther
GoogleOther
GoogleGeneric Google crawler for product teams fetching publicly accessible content (for example internal R&D); not tied to a single consumer productYesDataset / other
Storebot-Google
Storebot-Google
GoogleGoogle crawler for Google Shopping surfaces (Shopping tab, Google Shopping) with dedicated robots.txt token Storebot-GoogleYesDataset / other
FacebookExternalHit
facebookexternalhit
MetaFetches shared links to build previews (title, description, image) on Facebook, Instagram, Messenger, and related Meta appsPartialDataset / other
Meta-ExternalAds
meta-externalads
MetaCrawls the web for advertising and other business-related product and service improvementsYesDataset / other
Meta-ExternalAgent
meta-externalagent
MetaCrawls public web content for training foundation AI models or improving products by indexing content directlyYesTraining
Meta-ExternalFetcher
meta-externalfetcher
MetaFetches individual links at a user’s request and supports agentic AI product functions that may navigate websitesPartialLive browsing
Meta-WebIndexer
meta-webindexer
MetaCrawls the web to improve Meta AI search result quality and support citing and linking sources in responsesYesSearch indexing
ChatGPT-User
ChatGPT-User
OpenAIFetches pages when users browse the web during ChatGPT conversationsPartialLive browsing
GPTBot
GPTBot
OpenAICrawls public web content that may support OpenAI models and productsYesTraining
OAI-AdsBot
OAI-AdsBot
OpenAIValidates safety and relevance of landing pages submitted as ads on ChatGPT; not used to train foundation modelsYesDataset / other
OAI-SearchBot
OAI-SearchBot
OpenAIIndexes content for ChatGPT search / retrieval featuresYesSearch indexing
Perplexity-User
Perplexity-User
PerplexityUser-triggered fetches when Perplexity retrieves a page in-sessionPartialLive browsing
PerplexityBot
PerplexityBot
PerplexityCrawls and indexes pages for Perplexity search and answersYesSearch indexing

All bots remain in the page HTML for crawlers; filters only hide rows visually.

Which bots should I block or allow?

There is no universal “block all AI” or “allow everything” answer that fits every site. The useful approach is to match each bot’s category to your goals: AI visibility and citations, content licensing control, crawl-budget hygiene, or a mix of all three. Vinespire’s default editorial stance is pro-visibility for public marketing and documentation content, while staying factually neutral about the mechanics of each agent.

Training crawlers (for example GPTBot, ClaudeBot, Amazonbot, Bytespider, cohere-ai) collect public pages that may support model development. Allowing them can help future models know accurate facts about your brand; blocking them is appropriate when licensing, paywalls, or policy forbid training reuse. Remember that blocking one company’s training bot does not block another’s—and it does not remove content already present in older datasets.

Search-indexing bots (OAI-SearchBot, Claude-SearchBot, PerplexityBot) influence whether pages can appear in AI product search layers. Brands investing in generative engine optimization usually allow these agents and invest in clear, crawlable pages. A catch-all User-agent: * Disallow: / often blocks them by accident.

Live-browsing agents (ChatGPT-User, Claude-User, Perplexity-User) typically fetch pages when a person asks an assistant to open or use a URL. Volume is usually lower and demand-driven. Many sites allow them so conversational tools show accurate content; some operators document partial robots.txt applicability for user-fetch agents—read each bot’s official docs.

Dataset and product-token crawlers such as CCBot (Common Crawl), Diffbot, Google-Extended, and Applebot-Extended sit slightly outside the “one chat brand” mental model. CCBot feeds open corpora used by many downstream projects; Google-Extended and Applebot-Extended express preferences for generative product uses without necessarily equating to classic search crawlers. Knowing they exist matters even when they are less famous than GPTBot.

Practical workflow: inventory which agents you care about, write explicit robots.txt groups (avoid relying only on wildcards), deploy, then validate. robots.txt is not authentication—protect private content with real access controls.

Robots.txt AI Validator

Paste robots.txt and see which AI crawlers are allowed or blocked.

llms.txt Generator

Publish a curated site summary language models can retrieve at /llms.txt.

Free tools hub

More GEO utilities: readiness checks, schema generators, and estimators.

Glossary: AI Crawler · User-Agent · Training Data (glossary routes ship as the glossary hub expands).

Frequently asked questions

  • An AI crawler (or AI bot) is software that requests public web pages on behalf of an AI company or data project—often to train models, build search indexes for answer engines, extract structured data, or fetch a page when a user asks an assistant to open a URL. They are related to traditional search crawlers but serve different product goals and use distinct User-agent names in robots.txt.