{"pageUrl":"https://promagen.com/how-to-test-if-chatgpt-can-read-your-site","lastModified":"2026-05-10","provenanceHash":"sha256:d6a4e7221728d2d4fdebced3daa5ab9a6234e79a0b23bcaaec10eb6ca45a584d","provenanceNote":"OpenAI bot names (GPTBot, OAI-SearchBot, ChatGPT-User) are documented at platform.openai.com/docs/bots. Test methodology describes operational behaviour observable in ChatGPT's product UI as of 2026-05-10; no claim is made about ChatGPT's internal retrieval or ranking algorithms.","claims":[{"id":"claim-three-openai-bots","statement":"OpenAI documents three distinct user agents (GPTBot, OAI-SearchBot, ChatGPT-User) corresponding to model training, ChatGPT Search retrieval, and user-triggered fetches respectively. Server-log presence of these strings indicates the corresponding crawl behaviour.","evidenceUrl":"https://platform.openai.com/docs/bots","lastVerified":"2026-05-10","hash":"sha256:cd7dcf4dd2901843d17324f6fd4127d7641e03d3a6bb8245a12aae7adfb15199"},{"id":"claim-search-vs-model-knowledge","statement":"ChatGPT operates in two distinct retrieval modes: Search mode (live web fetch at answer time) and pure model knowledge (no fetch). The two produce different answers for the same query; testing should force the Search path to verify reachability.","evidenceUrl":"https://platform.openai.com/docs/bots","lastVerified":"2026-05-10","hash":"sha256:d8de35697a4828084bb6b4f25c268f5320e1278b8a521efb12d2310fd50a5d9d"},{"id":"claim-citation-rate-time-series","statement":"Citation rate per query, measured weekly against a fixed query set, is the appropriate operational signal for tracking ChatGPT visibility over time. Absolute citation counts without a denominator are less actionable than a per-query rate.","evidenceUrl":"https://promagen.com/sentinel/weekly","lastVerified":"2026-05-10","hash":"sha256:7f1dbbbca0b1a6618b6c3cb4eef48641d7fb89cd7491739220565341a1dce469"}]}