{
  "title": "Nice Pick AI Crawler Reality Report",
  "published": "2026-07-15",
  "publisher": "Nice Pick",
  "canonical_url": "https://nicepick.dev/research/ai-crawler-reality-report",
  "source": {
    "provider": "Cloudflare",
    "dataset": "HTTP Requests Adaptive Groups",
    "host": "nicepick.dev",
    "window": {
      "since": "2026-07-08T17:41:25.348Z",
      "until": "2026-07-15T17:41:24.348Z",
      "days": 7
    },
    "query_granularity": "One-day windows",
    "grouping": "User-agent string and HTTP response status"
  },
  "totals": {
    "requests": 513534,
    "http_2xx": 397088,
    "http_3xx": 66203,
    "http_4xx": 45017,
    "http_5xx": 5223,
    "http_5xx_rate": 0.0102
  },
  "classes": {
    "training_or_ingest": {
      "requests": 254595,
      "share_of_all_requests": 0.4958,
      "http_5xx": 4603
    },
    "retrieval_search_or_user_fetch": {
      "requests": 331,
      "share_of_all_requests": 0.0006,
      "http_5xx": 4
    },
    "search_engine": {
      "requests": 1926,
      "share_of_all_requests": 0.0038,
      "http_5xx": 3
    },
    "seo_tool": {
      "requests": 79189,
      "share_of_all_requests": 0.1542,
      "http_5xx": 142
    },
    "other_bot": {
      "requests": 266,
      "share_of_all_requests": 0.0005,
      "http_5xx": 0
    },
    "human_or_browser": {
      "requests": 177058,
      "share_of_all_requests": 0.3448,
      "http_5xx": 471
    }
  },
  "retrieval_actors": {
    "oai_searchbot": {
      "requests": 256,
      "http_5xx": 0
    },
    "chatgpt_user": {
      "requests": 44,
      "http_5xx": 0
    },
    "perplexitybot": {
      "requests": 30,
      "http_5xx": 4
    },
    "other_retrieval": {
      "requests": 1,
      "http_5xx": 0
    }
  },
  "derived": {
    "training_to_retrieval_ratio": 769.1691842900302,
    "training_requests_per_retrieval_request_rounded": 769
  },
  "citation_baseline": {
    "observed_on": "2026-07-03",
    "engine": "ChatGPT Search",
    "method": "One manual run of a fixed software-comparison prompt set with web search enabled",
    "queries": 10,
    "source_lists_containing_nicepick": 0,
    "note": "This baseline was measured separately and is not part of the July 8–15 Cloudflare request window. It is not a global or current citation-rate estimate."
  },
  "classification": {
    "training_or_ingest": "Known user agents associated with model training or bulk collection.",
    "retrieval_search_or_user_fetch": "Known user agents associated with search indexing, answer retrieval, or live user-directed fetches.",
    "note": "Classification is based on user-agent identity and apparent function. It does not prove how fetched content was ultimately used."
  },
  "limitations": [
    "Cloudflare HTTP Requests Adaptive Groups are analytics estimates, not raw access logs.",
    "User agents can be spoofed.",
    "Cloudflare daily grouped queries returned the 300 highest-volume user agents. Classified rows covered 513365 of 513534 requests; 169 long-tail requests fell outside those result sets.",
    "Requests are not unique pages, sessions, people, referrals, or citations. Assets and repeated fetches may contribute multiple requests.",
    "The human_or_browser class is unclassified browser-like traffic, not verified human traffic.",
    "A retrieval request does not prove that a page was cited in an answer.",
    "The 0/10 ChatGPT Search source-list baseline was a separate manual observation on 2026-07-03.",
    "This is one website over one week; the 769:1 request ratio should not be generalized to the web."
  ]
}
