ChatGPT Search visibility requires OAI-SearchBot access. OAI-SearchBot is used to show sites in search results, while GPTBot is a separate control used for foundation model training. Site owners can allow search while blocking training. WAF should look at not just user-agent but also OpenAI-published IP ranges.
OAI-SearchBot → ChatGPT search
GPTBot → foundation model training
ChatGPT-User → user-initiated fetch
robots.txt changes ≈ 24h
utm_source=chatgpt.comChatGPT Search visibility requires OAI-SearchBot access. OAI-SearchBot is used to show sites in search results, while GPTBot is a separate control used for foundation model training. Site owners can allow search while blocking training. WAF should look at not just user-agent but also OpenAI-published IP ranges.
Record brand, query, platform, source URL and measurement date in standard format.
Check robots, WAF, canonical, index, structured data and HTTP responses.
Do not interpret mention, citation, impression, referral and conversion values as the same metric.
Verify which hypothesis works with baseline, changelog, and remeasurement.
Do not use GPTBot setting for search visibility instead of OAI-SearchBot.
Meaning: The site may not be displayed in ChatGPT Search source answers.
Possible cause: robots.txt engeli.
Meaning: Content prefers not to be crawled for foundation model training.
Possible cause: Independent robot control.
Meaning: This is not an automatic crawler initiated by the user.
Possible cause: User-initiated ziyaret.
Meaning: OpenAI bot is allowed but still can't access the server.
Possible cause: Cloudflare or security rule.
Meaning: User-agent spoofed may be.
Possible cause: Bot spoofing.
Meaning: OpenAI systems may not update immediately.
Possible cause: Approximately 24-hour refresh.
Meaning: The blocked page may be found in some cases only by title/link.
Possible cause: Third-party index signal.
Meaning: ChatGPT referral URLs may contain measurement parameters.
Possible cause: OpenAI referral standard
No records matching this expression were found.
curl -sL https://example.com/robots.txt | grep -iE 'OAI-SearchBot|GPTBot|ChatGPT-User|User-agent|Allow|Disallow' | head -n 120Lists OpenAI bot rules.
curl -sSI -A 'OAI-SearchBot' https://example.com/rehber | grep -iE '^HTTP|location|x-robots-tag|content-type:'Displays the basic response for User-agent.
curl -sS https://openai.com/searchbot.json | python3 -m json.tool | head -n 80Displays published OAI-SearchBot IP ranges.
grep -Ei 'OAI-SearchBot|GPTBot|ChatGPT-User' access.log | tail -n 100Searches OpenAI user-agent requests in server logs.
grep -Ei 'chatgpt\.com|utm_source=chatgpt\.com' access.log | tail -n 100Searches for referral and UTM tracks.
curl -sL https://example.com/rehber | grep -ioE '<link[^>]+rel=["'"']canonical["'"'][^>]*>|<title>[^<]+' | head -n 5The source URL checks the title and canonical value.
User-agent: * Disallow: / to close everythingOAI-SearchBot and GPTBot manage their decisions in separate robots groups.OAI-SearchBot written for all requests allowVerify user-agent and OpenAI official IP range together.Count all visits to chatgpt.com with citationReferral session, source URL, conversion, and true citation are separated.SearchBot permission = strict ChatGPT citationPermissions only ensure appropriateness; they do not guarantee content and retrieval decisions.User-agent: OAI-SearchBot
Allow: /
User-agent: GPTBot
Disallow: /
User-agent: *
Allow: /
Sitemap: https://example.com/sitemap.xmlOAI-SearchBot and GPTBot robots settings are independent of each other.
Bot access may be blocked by the security layer even if robots are allowed.
ChatGPT referral traffic can be analyzed with UTM and referrer.
No. Technical feasibility, access, content quality, and measurement can be improved; citation and AI response depend on the dynamic systems of the platforms.
No. Crawl, index, canonical, content quality, and site reliability are the foundation of GEO work.
No. It is a community suggestion; Google Search explicitly states that it does not use it as a special visibility signal.
There is no special schema for Google AI features. Supported structured data must be used in a manner compatible with visible content.
No. Bot access only provides discovery and retrieval permission; source selection is not guaranteed.
AI impression, mention, citation, cited URL, referral, assisted conversion, and answer accuracy should be measured together.
There is no fixed duration. Crawl refresh, platform scope, content change, and user request vary accordingly.
We audit bot access, content, entity, citation and measurement processes for Google AI, ChatGPT Search, Bing Copilot and Perplexity. We do not guarantee exact citation or ranking.