ChatGPT-User vs Claude-SearchBot

ChatGPT-User and Claude-SearchBot compared on 9 documented fields, from 18 claims, last verified 2026-09-13.

Field	ChatGPT-User	Claude-SearchBot	Comparison
cites_sources	not_documented	not_documented	same
honours_crawl_delay	not_documented	yes	different
publishes_ip_list	yes	yes	same
purpose_documented	user_fetch	search	different
reads_llms_txt	not_documented	not_documented	same
respects_robots_txt	partial	yes	different
separate_search_and_training_tokens	not_documented	yes	different
user_agent_string_full	Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko); compatible; ChatGPT-User/1.0; +https://openai.com/bot	not_documented	different
verifiable_by_rdns	not_documented	not_documented	same

Differences

Field	ChatGPT-User statement	Claude-SearchBot statement
honours_crawl_delay	OpenAI does not publish whether ChatGPT-User honours the robots.txt Crawl-delay directive.	Claude-SearchBot respects the Crawl-delay directive in robots.txt when crawling domains.
purpose_documented	ChatGPT-User visits web pages for user actions in ChatGPT and in Custom GPTs.	Claude-SearchBot analyses online content to improve the relevance and accuracy of Claude search responses.
respects_robots_txt	ChatGPT-User fetches pages in response to user actions in ChatGPT, and OpenAI states that robots.txt rules may not apply to ChatGPT-User fetches.	Claude-SearchBot honours industry standard robots.txt directives that signal do not crawl.
separate_search_and_training_tokens	OpenAI does not document ChatGPT-User as a search or training opt-out token.	Anthropic separates search and training controls into two robots.txt tokens: Claude-SearchBot for search and ClaudeBot for training.
user_agent_string_full	ChatGPT-User sends the user agent string Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko); compatible; ChatGPT-User/1.0; +https://openai.com/bot.	Anthropic does not publish a full user agent string for Claude-SearchBot.