The AI bots, ranked by unwelcome

Every AI-associated user agent we check, and how many of the 645 readable sites block it. Snapshot: 2026-10-25

BotOperatorPurposeBlocked by
GPTBotOpenAItraining139 (21.6%)
ChatGPT-UserOpenAIfetch93 (14.4%)
OAI-SearchBotOpenAIsearch76 (11.8%)
ClaudeBotAnthropictraining144 (22.3%)
Claude-UserAnthropicfetch100 (15.5%)
Claude-SearchBotAnthropicsearch94 (14.6%)
Google-ExtendedGoogletraining120 (18.6%)
Applebot-ExtendedAppletraining125 (19.4%)
PerplexityBotPerplexitysearch118 (18.3%)
Perplexity-UserPerplexityfetch102 (15.8%)
CCBotCommon Crawltraining159 (24.7%)
BytespiderByteDancetraining162 (25.1%)
meta-externalagentMetatraining133 (20.6%)
AmazonbotAmazontraining119 (18.4%)
DuckAssistBotDuckDuckGosearch95 (14.7%)
MistralAI-UserMistralfetch95 (14.7%)
anthropic-aiAnthropictraining117 (18.1%)
Claude-WebAnthropicfetch117 (18.1%)
cohere-aiCoheretraining129 (20%)
cohere-training-data-crawlerCoheretraining98 (15.2%)
DiffbotDiffbottraining137 (21.2%)
omgiliWebz.iotraining130 (20.2%)
omgilibotWebz.iotraining114 (17.7%)
YouBotYou.comsearch114 (17.7%)
ImagesiftBotHivetraining114 (17.7%)
TimpibotTimpisearch112 (17.4%)
Meta-ExternalFetcherMetafetch95 (14.7%)
AI2BotAllen Institute for AItraining98 (15.2%)
DeepSeekBotDeepSeektraining79 (12.2%)
GrokBotxAItraining67 (10.4%)
PanguBotHuaweitraining94 (14.6%)
Gemini-Deep-ResearchGooglefetch71 (11%)

Purposes: training — collects data that may train models · search — builds an AI search index · fetch — retrieves a page live at a user's request. Sites are notably more willing to block training bots than fetch bots.