{
  "entity_a": {
    "slug": "claude-user",
    "name": "Claude-User"
  },
  "entity_b": {
    "slug": "gptbot",
    "name": "GPTBot"
  },
  "fields_documented": 11,
  "claims": 18,
  "last_verified": "2026-09-13",
  "fields": [
    {
      "field": "cites_sources",
      "value_a": "not_documented",
      "value_b": null,
      "claim_id_a": 44,
      "claim_id_b": null,
      "same": false
    },
    {
      "field": "honours_crawl_delay",
      "value_a": "yes",
      "value_b": "not_documented",
      "claim_id_a": 39,
      "claim_id_b": 79,
      "same": false
    },
    {
      "field": "publishes_ip_list",
      "value_a": "yes",
      "value_b": "yes",
      "claim_id_a": 40,
      "claim_id_b": 80,
      "same": true
    },
    {
      "field": "purpose_documented",
      "value_a": "user_fetch",
      "value_b": "training",
      "claim_id_a": 42,
      "claim_id_b": 82,
      "same": false
    },
    {
      "field": "reads_llms_txt",
      "value_a": "not_documented",
      "value_b": null,
      "claim_id_a": 45,
      "claim_id_b": null,
      "same": false
    },
    {
      "field": "respects_robots_txt",
      "value_a": "yes",
      "value_b": "yes",
      "claim_id_a": 37,
      "claim_id_b": 78,
      "same": true
    },
    {
      "field": "robots_txt_fetch_marker",
      "value_a": null,
      "value_b": "yes",
      "claim_id_a": null,
      "claim_id_b": 85,
      "same": false
    },
    {
      "field": "separate_search_and_training_tokens",
      "value_a": "not_documented",
      "value_b": "yes",
      "claim_id_a": 43,
      "claim_id_b": 83,
      "same": false
    },
    {
      "field": "user_agent_string_full",
      "value_a": "not_documented",
      "value_b": "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko); compatible; GPTBot/1.4; +https://openai.com/gptbot",
      "claim_id_a": 46,
      "claim_id_b": 84,
      "same": false
    },
    {
      "field": "user_fetch_exception_documented",
      "value_a": "no",
      "value_b": null,
      "claim_id_a": 38,
      "claim_id_b": null,
      "same": false
    },
    {
      "field": "verifiable_by_rdns",
      "value_a": "not_documented",
      "value_b": "not_documented",
      "claim_id_a": 41,
      "claim_id_b": 81,
      "same": true
    }
  ],
  "differences": [
    {
      "field": "cites_sources",
      "statement_a": "Anthropic does not publish whether Claude responses cite pages fetched by Claude-User.",
      "statement_b": null,
      "claim_id_a": 44,
      "claim_id_b": null
    },
    {
      "field": "honours_crawl_delay",
      "statement_a": "Claude-User respects the Crawl-delay directive in robots.txt when fetching pages.",
      "statement_b": "OpenAI does not publish whether GPTBot honours the robots.txt Crawl-delay directive.",
      "claim_id_a": 39,
      "claim_id_b": 79
    },
    {
      "field": "purpose_documented",
      "statement_a": "Claude-User accesses websites in response to questions that individual users ask Claude.",
      "statement_b": "GPTBot crawls content that may be used to train OpenAI generative AI foundation models.",
      "claim_id_a": 42,
      "claim_id_b": 82
    },
    {
      "field": "reads_llms_txt",
      "statement_a": "Anthropic does not publish whether Claude-User reads llms.txt files.",
      "statement_b": null,
      "claim_id_a": 45,
      "claim_id_b": null
    },
    {
      "field": "robots_txt_fetch_marker",
      "statement_a": null,
      "statement_b": "OpenAI may add a robots.txt marker to the GPTBot user agent string when GPTBot fetches robots.txt files.",
      "claim_id_a": null,
      "claim_id_b": 85
    },
    {
      "field": "separate_search_and_training_tokens",
      "statement_a": "Anthropic does not document Claude-User as a search or training opt-out token.",
      "statement_b": "OpenAI separates search and training controls into two robots.txt tokens: OAI-SearchBot for search and GPTBot for training.",
      "claim_id_a": 43,
      "claim_id_b": 83
    },
    {
      "field": "user_agent_string_full",
      "statement_a": "Anthropic does not publish a full user agent string for Claude-User.",
      "statement_b": "GPTBot sends the user agent string Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko); compatible; GPTBot/1.4; +https://openai.com/gptbot.",
      "claim_id_a": 46,
      "claim_id_b": 84
    },
    {
      "field": "user_fetch_exception_documented",
      "statement_a": "Anthropic applies the same robots.txt statement to Claude-User as to ClaudeBot and Claude-SearchBot and documents no exception for user-initiated fetches.",
      "statement_b": null,
      "claim_id_a": 38,
      "claim_id_b": null
    }
  ],
  "sources": [
    "https://claude.com/crawling/bots.json",
    "https://developers.openai.com/api/docs/bots",
    "https://openai.com/gptbot.json",
    "https://support.claude.com/en/articles/8896518-does-anthropic-crawl-content-from-the-web-and-how-can-site-owners-block-the-crawler"
  ]
}