Crawler-labelled requests.

A recent sample of up to 10,000 requests matching 25 AI and search user-agent labels. These labels are supplied by the requester; they do not verify who made the request. This log stores no IP addresses; this view shows aggregates only.

10,000Matching GET requestsin this recent sample
20 of 25User-agent labels matchedin this recent sample
332Distinct paths requestedin this recent sample
Sample totals, most requests first – (UTC)
User-agent labelRequestsPathsLatest in sample
meta-externalagentMeta8,381169
OAI-SearchBotOpenAI1,230260
PerplexityBotPerplexity162125
GooglebotGoogle5114
AmazonbotAmazon4949
BingbotMicrosoft2414
ApplebotApple2115
Claude-UserAnthropic168
ChatGPT-UserOpenAI127
BytespiderByteDance108
DuckAssistBotDuckDuckGo96
GPTBotOpenAI76
CCBotCommon Crawl55
Google-ExtendedGoogle55
cohere-aiCohere55
ClaudeBotAnthropic44
Perplexity-UserPerplexity33
Claude-SearchBotAnthropic22
MistralAI-UserMistral22
YouBotYou.com22

Totals cover the latest 10,000 matching records available at the last successful read. Older records may fall outside this sample.

Earliest match in this sample
  1. meta-externalagent
  2. Bingbot
  3. Amazonbot
  4. Googlebot
  5. Bytespider
  6. OAI-SearchBot
  7. PerplexityBot
  8. Applebot
  9. ChatGPT-User
  10. GPTBot
  11. Claude-User
  12. DuckAssistBot
  13. Claude-SearchBot
  14. Perplexity-User
  15. ClaudeBot
  16. Google-Extended
  17. CCBot
  18. cohere-ai
  19. YouBot
  20. MistralAI-User

The earliest request for each user-agent label in this recent sample, not its first-ever match.

No recorded matches in this sample (5)
  • Claude-Web · Anthropic
  • anthropic-ai · Anthropic
  • FacebookBot · Meta
  • Applebot-Extended · Apple
  • cohere-training-data-crawler · Cohere

These records are written before the downstream response. They do not show whether content was retrieved successfully, indexed, cited or recommended. A user-agent label can also be supplied by a test or another client. Successful reads refresh hourly; failed reads are retried on a later request.