How to Get Cited by Perplexity AI, Not Just ChatGPT

To get cited by Perplexity AI, your page has to survive a live fetch. Perplexity sends its own crawler, PerplexityBot, to read your page in real time when a person asks a question, then selects the sources that add information the other results do not already contain. Be reachable, be fast, and say something new.

Claude works differently. Claude leans on the Brave Search index plus its own training corpus, not a live crawler run the way Perplexity runs PerplexityBot. Getting cited by both engines, not just ChatGPT, means being crawlable, fast, formatted for machine extraction, and genuinely more useful than the other ten pages competing for the same answer.

• Perplexity: reads your live page with PerplexityBot, so speed and crawlability decide whether you are even in the running.
• Claude: pulls from the Brave Search index and its training data, so search presence and outside mentions matter more than raw page speed.

Here is a number almost nobody publishes. Across the 8 business sites No Time For SEO tracks in raw server logs, AI crawlers and agents hit us 28,994 times in one week, the week of Monday July 13. PerplexityBot accounted for 1,190 of those fetches, ClaudeBot 2,393, and GPTBot 1,631.

Takeaway: live answer-time fetches from chatgpt-user outnumber every training crawler, so your page has to be reachable at the exact moment a question is asked.

Full disclosure in one sentence: No Time For SEO manages AI search and organic search for a small set of businesses, which is why these server logs exist, and every step in this guide is something you can do yourself.

How does Perplexity actually choose which pages to cite?

Perplexity AI chooses citations by sending its own crawler, PerplexityBot, to fetch candidate pages in real time, then ranking them by relevance and information gain rather than by Google position. A top-10 Google ranking is not a ticket to a Perplexity citation.

1. Split the question into several sub-queries.
2. Retrieve candidate URLs from Perplexity’s own live index and partners.
3. Fetch the promising pages on the spot with PerplexityBot.
4. Drop any page that is too slow or blocked.
5. Assemble the answer from the sources that each add something distinct.

This is why generic explainer pages lose. If your page repeats the same five tips every other agency blog repeats, Perplexity’s source-selection step has no reason to cite you over the page it already picked. Perplexity optimizes for coverage of the answer, not for rewarding your effort.

Nobody outside Perplexity knows the exact ranking weights, and anyone who claims to is guessing. What No Time For SEO can measure is behavior: which bots hit our pages, how fast the pages responded, and which pages later show up as citations. Everything here is grounded in that observed behavior.

How is getting cited by Claude different from getting cited by Perplexity?

Claude reaches your page mainly through the Brave Search index and its own training corpus, not through a live crawler that Anthropic runs the way Perplexity runs PerplexityBot. So the levers for Claude differ from the levers for Perplexity, and treating all AI search as one thing is the most common mistake.

• Perplexity lever: be fast and crawlable at answer time, because Perplexity fetches you live.
• Claude lever: be visible in Brave Search and mentioned widely enough that the model has already seen you.

Here is the side-by-side that no ranking page on this query lays out cleanly. Read it as two separate pipelines you optimize in parallel.

Takeaway: optimize Perplexity for speed and crawlability, and optimize Claude for search presence and being cited elsewhere.

How often does ClaudeBot crawl compared with PerplexityBot?

ClaudeBot hit our tracked sites 2,393 times that week, nearly double PerplexityBot’s 1,190. ClaudeBot mostly gathers training and search data rather than answering a live question, so a ClaudeBot visit is Anthropic building its picture of you, not a person waiting on an answer right now.

Why does Perplexity ignore my page and cite Wikipedia instead?

Perplexity cites Wikipedia instead of your page when your page repeats what Wikipedia already says. Perplexity rewards information gain, meaning content that adds facts the existing sources lack, so a page that only restates common knowledge gives the engine no reason to add it.

Wikipedia wins the generic-summary slot almost every time. It is comprehensive, heavily referenced, and already trusted. You will not out-Wikipedia Wikipedia. You beat it by publishing something Wikipedia structurally cannot.

• Original data from your own logs, tools, or sales records.
• A real case study with named steps and measured outcomes.
• A specific number: a price, a timeline, a first-hand result.

A concrete example of information gain: a single researched article No Time For SEO wrote for a 5-location dental group earned 52,171 Google impressions in one month while ranking on page one, measured July 2026 from Google Search Console. No competitor can copy those numbers.

Takeaway: the citable part is the measured result, not the advice around it, because the advice already sits on a hundred other pages.

So what is information gain, in one line?

Information gain is the amount of new, non-duplicated information your page adds beyond what the already-cited sources contain. Perplexity needs it because its job is to cover an answer with the fewest, most complementary sources, so a page that adds nothing new earns nothing.

How do I stop Cloudflare from blocking PerplexityBot and ClaudeBot?

Cloudflare blocks PerplexityBot and ClaudeBot by default on every zone created after July 2025 through its managed AI-bot rule, so on a newer site you must explicitly allow them or they never reach your page. This is the most common reason a technically good page is invisible to Perplexity.

Two separate things can block a bot, and people confuse them. Your robots.txt is a polite request that compliant bots obey. Cloudflare’s managed rule is an enforced block at the edge that stops the request cold. You have to check both.

Here is exactly how to open the door for the AI crawlers you want, inside your Cloudflare dashboard.

1. Log in to your Cloudflare dashboard and select the site’s zone.
2. Open Security, then Bots (on older layouts it lives under Security, then WAF).
3. Find the managed rule labeled Block AI bots or Block AI Scrapers and Crawlers.
4. Set it to allow the verified bots you want cited by, or switch the managed block off to let all compliant AI crawlers through.
5. Check your robots.txt separately and confirm it does not Disallow PerplexityBot or ClaudeBot.
6. Verify with a curl test using each bot’s user-agent, then watch your server logs for real hits.

You should also understand the difference between being blocked and being delisted, because they call for opposite fixes.

• Blocked: your edge stops PerplexityBot, so the page is never read.
• Delisted: the engine chooses to drop a page it can reach, usually over a crawling-policy fight, which is what Cloudflare and Perplexity clashed over in the August 2025 stealth-crawling dispute.

For most site owners the decision is simple. If you want AI citations, allow the declared, verifiable bots, and block only the ones that ignore your rules and identify themselves falsely.

Does Perplexity cite pages that load slowly?

Perplexity usually will not cite pages that load slowly, because its real-time fetch has a short timeout window before it moves on to a faster competing source. When you get dropped for speed, there is no error and no warning; you simply are not in the answer.

What is a 499 error, and why does it matter for AI citations?

A 499 is the status a server records when the client closes the connection before the response finishes. If Perplexity’s fetcher gives up on a slow page, that abandoned request is the answer-engine version of a bounce. The practical target is a Time to First Byte under 500 milliseconds on the pages you want cited, fixed in this order.

1. Pillar pages first, since they carry the most answers Perplexity can pull.
2. FAQ pages next, because they map directly to how people phrase questions.
3. Money pages last, so a live fetch never times out on a page that converts.

Claude is less exposed to this specific problem, because Claude often relies on the Brave index and its training data rather than a live fetch with a hard timeout. That is one more reason to treat the two engines separately: speed is a Perplexity-critical lever and a Claude-secondary one.

What content format does Perplexity prefer to cite?

Perplexity prefers to cite content that leads with a direct, quotable answer in the first 100 words, uses question-based headings that mirror how people ask, and formats facts as bullet lists, numbered steps, and comparison tables that are easy to extract. Buried answers get skipped.

• Lead with the answer in the first 100 words, not after three intro paragraphs.
• Use question-phrased headings that match natural-language queries.
• Format facts as bullet lists, numbered steps, and comparison tables.

Schema matters too. Schema-enabled pages see meaningfully higher citation rates across AI answer engines, so mark up your content with the right JSON-LD types. Use these as your default set.

Takeaway: match the schema type to the visible structure of the page, and never fake schema for content that is not actually there.

Which formatting rules help across every engine?

Write question-phrased H2 headings, keep each paragraph to one idea, put the most important sentence first, and add a comparison table wherever you weigh two options. This is how retrieval systems chunk and quote a page. For the deeper mechanics, see our guide on how AI chatbots choose which business to recommend.

Does content freshness change citation odds?

Both Perplexity and Claude weight recency, and Perplexity in particular works within short recrawl windows, so updating a strong page on a regular cadence and showing a visible published and updated date keeps it eligible. Build author credentials and in-content citations too; these E-E-A-T signals are how you earn entity authority over time.

How do I verify a crawler can actually reach and cite my page?

You can verify a crawler reached your page by checking your raw server logs for PerplexityBot and ClaudeBot user-agents, or by running a curl test with each bot’s user-agent and confirming a 200 response. This step is the one almost every guide skips, and it is the only way to know your optimization is not wasted on a page the bot never sees.

Run these four checks in order. They take about ten minutes and turn guesswork into evidence.

1. Grep your access log for the user-agents: run grep -Ei ‘PerplexityBot|ClaudeBot|GPTBot|chatgpt-user’ on your nginx or Apache access log and read the hits.
2. Fetch the page as the bot: run curl -A ‘PerplexityBot’ -I https://yourdomain.com/your-page/ and confirm the status is 200, not 403 or 429.
3. Open Cloudflare’s Crawl Control view and read the per-bot request and block log.
4. Time the response: run curl -w ‘TTFB: %{time_starttransfer}s’ -o /dev/null -s https://yourdomain.com/your-page/ and check it is under half a second.

A 403 or 429 in the fetch test means something is blocking the bot, usually Cloudflare or a firewall rule, so return to the Cloudflare section. A clean 200 plus real PerplexityBot lines in your log is proof the door is open. For why a business can be invisible to these systems, read why ChatGPT doesn’t know your business exists.

What does a live chatgpt-user hit tell me?

A chatgpt-user hit is a live answer-time fetch. In the week we measured, 6,370 fetches came from chatgpt-user, the agent ChatGPT uses when it reads a page live to answer a real person. Seeing chatgpt-user land on a URL in your logs means a human’s question just pulled your page into an answer, the closest thing to a real-time citation signal you will get.

How do I monitor whether I have actually been cited?

You can monitor AI citations after publishing by asking each engine your target questions on a fixed schedule and logging which URLs it cites, because none of the engines send you a referral notice the way analytics logs a click. If you do not track it deliberately, you never learn whether the work paid off.

1. List the exact questions a buyer would ask, including the QFO sub-queries you targeted.
2. Ask Perplexity, Claude, and ChatGPT those questions on a fixed schedule; weekly is enough.
3. Record which URLs each engine cites and whether yours appears, in a simple spreadsheet.
4. Watch your server logs for chatgpt-user and PerplexityBot fetches hitting the URL right after you publish.

Cross-engine citation is the signal to watch. Appearing in Perplexity, Claude, and ChatGPT for the same query correlates with higher perceived quality, because each engine amplifies existing citation networks rather than judging you alone. One citation is luck; three across engines is a pattern.

How soon should I expect a Perplexity or Claude citation?

Perplexity fetches live, so a crawlable fast page can appear quickly, while Claude often lags because it leans on Brave recrawl and training updates. No Time For SEO never promises a certain citation or ranking, and anyone promising a locked-in position is overselling. Track weekly, then again after 30 days.

What does non-branded strength have to do with AI citations?

A trading-education platform No Time For SEO runs search for has 82 of its top 100 keywords on page one of Google, and 63% of its clicks are non-branded, measured July 2026 across the GSC top-100 query sample. Non-branded strength is the demand signal that feeds AI citation over time. For the ChatGPT side of this playbook, see how to get your business mentioned in ChatGPT.

Your do-this-now checklist for this week

Work this list top to bottom. Each step is something a busy owner can finish this week, and each one removes a specific reason Perplexity or Claude is not citing you yet.

1. Allow PerplexityBot and ClaudeBot in Cloudflare and confirm robots.txt does not disallow them.
2. Curl-test your top page as PerplexityBot and confirm a 200 response, not a 403.
3. Grep your server logs to prove the AI bots are actually reaching the page.
4. Measure TTFB and get your money pages under 500 milliseconds.
5. Rewrite the first 100 words of each target page into a direct, quotable answer.
6. Add one original number, table, or case result that no competing page has.
7. Mark up the page with FAQPage, HowTo, or Article schema that matches the visible content.
8. Log a weekly citation check across Perplexity, Claude, and ChatGPT so you can see progress.

Do these eight and you have addressed every mechanical reason a good page stays uncited: blocked at the edge, too slow to fetch, a buried answer, nothing new to add, no schema, and no way to know if it worked.

Frequently asked questions

What content format does Perplexity prefer to cite?

Perplexity prefers a direct answer in the first 100 words, question-based headings, bullet lists, numbered steps, and comparison tables, backed by FAQPage, HowTo, or Article schema. Buried or unstructured answers get skipped because the engine cannot cleanly extract and quote them.

How do I stop Cloudflare from blocking PerplexityBot?

Open your Cloudflare dashboard, go to Security then Bots, find the managed Block AI bots rule, and set it to allow verified crawlers or turn the block off. Then confirm robots.txt does not disallow PerplexityBot, and curl-test the page to check for a 200 response.

Does Perplexity cite pages that load slowly?

Perplexity usually will not cite slow pages, because its real-time fetch has a short timeout before it moves to a faster source. Aim for a Time to First Byte under 500 milliseconds on the pages you want cited, prioritizing pillar, FAQ, and money pages first.

How is getting cited by Claude different from Perplexity?

Claude leans on the Brave Search index and its training corpus, while Perplexity runs its own live crawler, PerplexityBot, and fetches pages in real time. So Perplexity rewards speed and crawlability, and Claude rewards Brave visibility and earned third-party mentions.

Why does Perplexity ignore my page and cite Wikipedia instead?

Perplexity cites Wikipedia when your page only restates common knowledge Wikipedia already covers. To win a slot, publish original data, a case study, or a specific measured result, because Perplexity selects complementary sources by information gain, not by effort.

What is information gain and why does Perplexity need it to cite me?

Information gain is the new, non-duplicated information your page adds beyond the already-cited sources. Perplexity needs it because its goal is to cover an answer with the fewest complementary sources, so a page that adds nothing new gives the engine no reason to name it.

How long until Perplexity or Claude cites a new page?

There is no fixed timeline, and nobody outside the engines can promise one. Because Perplexity fetches live, a crawlable fast page can appear quickly; Claude often lags because it depends on Brave recrawl and training updates. Monitor weekly rather than expecting an instant result.