Learn how to get cited by Perplexity: allow its crawler, state product facts plainly, and publish the comparison pages its answers need.
Let PerplexityBot fetch a clear page, keep the important facts in plain text, and publish the comparison pages people actually search for. That is the direct answer.
The rest comes down to where Perplexity looks, what it can extract, and whether the cited page still supports the answer. In the 1,787 Perplexity answers we sampled, it cited 5.3 sources per answer on average and named the scanned product in 8.2% of answers.
Perplexity's own bot documentation says:
“PerplexityBot is designed to surface and link websites in search results on Perplexity.”
Attributed to Perplexity's bot documentation.
That sentence sets the job description. PerplexityBot is the crawler used for discovery and search results. The same documentation says it is not used to crawl content for AI foundation models. That distinction matters here: what follows is about search access and citations, not about whether your content trains a model.
Perplexity recommends allowing PerplexityBot in your robots.txt file and permitting requests from its published IP ranges. Its full user agent is Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; PerplexityBot/1.0; +https://perplexity.ai/perplexitybot).
That advice still applies to indexing and discovery. Your robots.txt should not block PerplexityBot, and your WAF should recognise its current IP ranges, which Perplexity updates regularly. PerplexityBot and robots.txt covers the rule and how to check it.
There is a second mechanism. When a user's query sends Perplexity to a specific page during an answer, an on-demand fetcher may visit that page. Perplexity says this user-requested fetch generally ignores robots.txt because a person prompted it. That does not make robots.txt irrelevant for the PerplexityBot crawler or make your site reliably discoverable.
Use the free AI readiness check to inspect the wider set of access and page checks. In our audit of 120 sites, the AI crawler check passed on 115 sites and failed on 5. Access is common, but it is not universal.
Make the page fetchable before you spend time polishing its authority signals.
A crawler can reach a page and still find no useful answer. State what the product is, who it is for, what it costs, and what it is being compared with in visible text, rather than leaving those facts to a script, an image, or a slogan.
In our audit of 120 sites, 118 homepages passed a check for answering a plain request. The check for a plain statement of what the product is, for whom, at what price, and against what passed on 20 sites, failed on 43 and warned on 55.
Buyers ask for a fit, a cost, or an alternative. Put those answers where a reader and a fetcher can find them.
The title and description check passed on 92 of 120 sites, while a clear H1 passed on 100. The page still needs to answer the buying question in ordinary language.
Here is what that can look like on one unnamed product page.
Before: the H1 read “Work better, together”, the first three paragraphs described company values, and the price appeared only after a “Contact us” link several screens down. The feature grid named integrations but never said which teams the product served.
After: the H1 said what the product was for, the opening sentence named its intended team, and a pricing link sat beside the opening explanation. A section below compared it with the main alternative and stated when each option made sense.
The page still had room for detail. Readers no longer had to assemble the answer themselves.
Use this checklist when you edit a page:
Price was visible or linked on 90 of 120 sites. Content was present without JavaScript on 91, with 18 failures and 9 warnings. A sitemap was published on 99, while canonical and Open Graph tags passed on 91 and warned on 27.
Metadata helps Perplexity identify a page. It cannot supply the missing product statement.
Perplexity does not need to cite only your own site. In the recorded citation data, listicles and roundups accounted for 67.5%, or 30,585 citations. Competitor product pages accounted for 24.4%, or 11,052 citations.
Together, those page types accounted for 91.9% of citations in the dataset. Community threads accounted for 3.4%, or 1,534 citations. The scanned product's own site accounted for 1.5%, or 663 citations.
That changes where you look for opportunities. Your product page can explain the offer, while a roundup or comparison page may explain the relationship between your product and the buyer's question more directly.
Build a map of the questions you want to answer. Search for the listicles, roundups and competitor pages that already receive citations for those questions. Check whether the facts about your product are current and whether the page can state them without hedging.
You do not need to copy those pages. You do need to understand the page type Perplexity already cites for that kind of answer. The weekly research on who owns AI answers gives you that wider view, while the CiteGraph methodology explains how the citation sample is collected.
Your own site still matters, but its role is smaller in this dataset: it supplied 1.5% of recorded citations, or 663 citations. Treat it as one part of the citation path, not the whole path.
A Perplexity-friendly page puts a clear answer near the top and supporting detail below it. The answer should survive a quick fetch, while the detail helps a reader decide whether the product fits.
A pricing page should make price and limits easy to find. A comparison page should name both options and explain the difference. The same logic applies whether the page is about alternatives or documentation: the answer has to be visible before the reader has to dig for it.
Use headings that name the question or subject. Keep related facts together, and put definitions, limits and dates beside the claims they qualify.
The readiness audit found that questions were answered on the page on 56 of 120 sites. It warned on 62 and failed on none. Comparison or alternatives pages passed on 54, failed on 20 and warned on 46.
Schema can help with entity and product context, but it cannot replace prose. Organization schema passed on 72 of 120 sites. Product or SoftwareApplication schema passed on 60 and failed on 58.
Keep the canonical version reachable, render the main content without JavaScript, and avoid competing copies of the same page. Update price, availability and product scope where they change.
Citations are an outcome, so measure the outcome. Check whether Perplexity names the product, which URL it cites, what page type supplied the citation, and whether the cited page still says what the answer claims.
As noted above, our Perplexity sample named the scanned product in 8.2% of answers. For comparison, Gemini named it in 6.8% of 1,784 answers, the ChatGPT family in 4.4% of 1,781, and Claude in 6% of 1,467.
More cited sources create more possible entry points. They also make the result harder to reduce to a single ranking position. Your product may be named without your site being cited, or your site may be cited without the product being named.
Look for changes after useful page edits, access fixes and updates to comparison content. Do not treat one answer as a verdict, because Perplexity can choose different sources for different wording, users and fetches.
The free AI visibility checker can help you inspect where your product appears. Use it alongside server logs, crawler checks and manual review of the cited pages. A citation that points to an old price or a thin page is a signal to fix the source.
Our data has limits. We measure samples, not every Perplexity answer, and the citation shares come from the pages and queries in those samples. We do not know what we have not measured.
The opening answer has not changed: let Perplexity fetch the page, state the product facts plainly, and publish the comparison material its answers need. What has changed is the route you can inspect: crawler access for discovery, page content for extraction, and the external pages that supply most recorded citations.
Make the page fetchable, state what the product is and who it serves, and keep important facts in visible text. Perplexity can choose different sources for different wording, users and fetches.
Perplexity recommends allowing PerplexityBot in robots.txt and permitting requests from its published IP ranges. An on demand fetch triggered by a user's request generally ignores robots.txt, but normal discovery still depends on crawler access.
Put a clear answer near the top and state the product, audience, price and comparison in ordinary language. Keep the main answer available without JavaScript and place supporting detail below it.
In the recorded citation data, listicles and roundups accounted for 67.5% of citations, while competitor product pages accounted for 24.4%. Together, those page types supplied 91.9% of citations in the dataset.
Check whether Perplexity names the product, which URL it cites, what page type supplied the citation, and whether the cited page supports the claim. Review changes after access fixes, page edits and comparison content updates.
Cite this: CiteGraph, “How to get cited by Perplexity”, 11 Sept 2026, https://www.citegraph.app/blog/how-to-get-cited-by-perplexity