ChatGPT Search Updates
What ChatGPT Search’s evolving citations and crawler controls mean for website owners.
What matters now
ChatGPT Search can use web results to produce current answers and may show inline citations or a Sources panel. For website owners, the practical takeaway is simple: make important public pages accessible, clear, accurate, and useful enough to support a real answer.
What ChatGPT Search can show
When ChatGPT uses search, it may generate an answer informed by current web sources. Depending on the interface and query, you may see inline citations beside individual claims, or a Sources control below the answer that opens a list of cited and related links.
Inline citations
Small citation links can appear next to statements in a search response. Users can inspect them to see which source supported that part of the answer.
Sources panel
When inline citations are not visible, ChatGPT may provide a Sources control with cited pages and other relevant links.
The crawler distinction that matters
Not every OpenAI crawler has the same purpose. If your goal is to make public content eligible to appear in ChatGPT Search results, the relevant user agent is OAI-SearchBot, not necessarily GPTBot.
OAI-SearchBot
OpenAI identifies OAI-SearchBot as the crawler relevant to inclusion in ChatGPT Search summaries and snippets. If it is blocked in robots.txt, your pages may not be eligible to appear in those search experiences.
GPTBot
GPTBot is associated with collecting public content that may be used to improve OpenAI foundation models. Allowing or blocking GPTBot is a separate choice from allowing ChatGPT Search access.
ChatGPT-User
ChatGPT-User can be used for user-initiated actions, such as a user asking ChatGPT to visit a specific page. It is not the primary control for general ChatGPT Search visibility.
A practical robots.txt check
If you want your public pages to remain eligible for ChatGPT Search, verify that you are not unintentionally blocking OAI-SearchBot. Keep sensitive areas—such as dashboards, staging environments, and account pages—protected separately.
# Example: allow ChatGPT Search crawling on public content User-agent: OAI-SearchBot Allow: / # Optional: choose training access separately User-agent: GPTBot Disallow: / # Keep sensitive areas restricted User-agent: * Disallow: /admin/ Disallow: /dashboard/ Disallow: /account/
This is only an example. Review your own privacy, licensing, security, and content-distribution requirements before changing crawler rules.
What to improve beyond crawler access
Allowing a crawler removes one access barrier. It does not guarantee that a page will be indexed, retrieved, cited, or recommended. Focus on the signals you can control:
- Give key pages a direct answer near the top, followed by useful detail and descriptive headings.
- Keep product, pricing, author, and company information accurate and current.
- Make pages easy to crawl with meaningful HTML, internal links, canonical URLs, and a clean sitemap.
- Use structured data to reinforce the meaning of visible content, not as a shortcut to guaranteed citations.
- Track a consistent set of relevant prompts and look for trends over time rather than relying on one answer.
What to monitor
Check your AI search foundation
GazeRank AI can help you review crawlability, content structure, trust signals, schema, and other factors that affect how clearly your website can be understood by AI search systems.
Related learning