Why does Perplexity cite my site but ChatGPT doesn't?

Each AI engine builds answers from its own pipeline. Your site has to clear every stage of that pipeline separately: the crawler must be allowed, the network must let it through, the page must sit in that engine's index, and the engine must decide to search for the question at all.

Perplexity and ChatGPT run different versions of each stage:

• Crawler. Perplexity indexes with PerplexityBot. ChatGPT search uses OAI-SearchBot. A robots.txt rule for one does nothing for the other.
• Index and providers. Perplexity searches its own index. OpenAI's help page says ChatGPT search rewrites your question into one or more targeted queries and sometimes partners with other search providers.
• Whether search runs. ChatGPT searches at the model's discretion. When it doesn't, the answer comes from what the model learned in training, and no live page gets cited.

So a Perplexity citation tells you PerplexityBot reached your page and Perplexity's retrieval picked it. It tells you nothing about OpenAI's crawler, your firewall's treatment of OpenAI's IP ranges, or whether ChatGPT searched. Treat each engine as its own audit.

As of October 2026, OpenAI's crawler documentation describes three bots with different jobs:

• GPTBot governs training use only. Allowing it does not put you in ChatGPT search.
• OAI-SearchBot controls ChatGPT search. OpenAI says opted-out sites "will not be shown in ChatGPT search answers, though can still appear as navigational links."
• ChatGPT-User handles fetches a user triggers, and OpenAI says robots.txt rules "may not apply" to it.

OpenAI states each setting is independent. A common error is a blanket AI block, written to stop training, that also removes ChatGPT search eligibility while an older rule leaves Perplexity open.

The second gate is the network. OpenAI's help center says eligibility means allowing OAI-SearchBot and confirming your host or CDN allows traffic from OpenAI's published search bot IP addresses. A firewall can return 403 to OAI-SearchBot while letting PerplexityBot through, and robots.txt will never show it.

Defaults can change without you editing anything. From 15 September 2026, Cloudflare blocks its Training and Agent crawler categories by default on ad-displaying pages for new domains, new sites on existing accounts and existing free-tier customers, while Search stays allowed. Check your CDN's per-category settings instead of assuming. If robots.txt already looks right, work through what else blocks ChatGPT when robots.txt allows AI bots.

OtterlyAI published a 2026 test of two new sites, each loaded with about 1,000 AI-generated posts, tracked across seven engines for about four months.

Otterly tracked 3,934 citations for one site, capybara-world.com, and 2,718 of them came from Microsoft Copilot. Perplexity and Claude cited the sites too. ChatGPT citations were effectively zero, even though ChatGPT's crawlers kept visiting and the sites ranked in Bing.

Two lessons follow:

• One engine's citations predict nothing about another's. A single blended score would have hidden the split entirely.
• Crawler visits are not citations. ChatGPT fetched pages it never cited, so a bot in your logs is not proof of inclusion.

The same experiment explains why Copilot can cite a site that Google's AI features ignore, seen from the index side.

Keep the limits in view. It is one pair of sites on a low-stakes topic, the prompt list is not public, and a tracking vendor ran it. Read it as a clear example of divergence, not a measured rate.

ChatGPT looked at you but didn't cite you

OpenAI's API returns the sources a search consulted separately from the ones it cited, and says the consulted set is often larger. Retrieval is not citation. The guide to cited versus retrieved sources explains the gap.

ChatGPT didn't search

Without search, ChatGPT answers from training data frozen at a cutoff date. Nothing you publish this quarter reaches that answer. If the answer shows no links, assume search did not run.

Your Perplexity citation came from a user fetch

Perplexity says Perplexity-User "generally ignores robots.txt rules." A citation can appear even where PerplexityBot is blocked, so it may not prove what you think. See how PerplexityBot and Perplexity-User differ.

Content only appears after JavaScript

Vercel found ChatGPT's and Claude's crawlers fetch JavaScript files but don't execute them. If the fact ChatGPT needs loads client-side, it may be invisible.

ChatGPT cites other sites instead

When an engine answers from third-party pages, being named on those pages is associated with appearing in the answer. Start by finding the pages AI engines cite.

1. Read robots.txt per bot. Confirm OAI-SearchBot is allowed. Decide on GPTBot separately, since it only governs training.
2. Request a page as OAI-SearchBot. Look at the status code. A 403, 429 or challenge page means your firewall or CDN is turning it away. The AI Crawler Checker requests a page as each AI crawler and shows which ones get blocked.
3. Check index status engine by engine. Use the per-engine index status playbook. A healthy index in one engine tells you nothing about the next.
4. Put key facts in the raw HTML. Prices, product names and one-line definitions should be present before JavaScript runs.
5. Re-test the same buyer questions in ChatGPT. Note whether the answer shows sources, which tells you whether search ran.

None of this guarantees ChatGPT will cite you. Ranking and selection stay with OpenAI. What these steps do is remove the blocks that are yours to remove.