How Do You Get Cited by Perplexity? Crawler Access, Numbered Sources and Narrow Answers

To get cited by Perplexity, make sure PerplexityBot can crawl your pages, then make each page the clearest and most current answer to one specific question. Access is a yes-or-no gate; citability is a competition you win one question at a time.

Perplexity's crawler documentation says PerplexityBot is designed to surface and link websites in search results on Perplexity, and that it is not used to crawl content for AI foundation models. To appear in results, Perplexity recommends allowing PerplexityBot in robots.txt and permitting requests from its published IP ranges. Once you are in its index, the question becomes which page best supports each sentence of an answer, and that is where narrow, specific pages beat broad ones.

If you are working on several engines at once, our step-by-step AEO process gives you the shared sequence; this page covers what is particular to Perplexity.

Perplexity builds answers from web search results and links each claim back to a source. Its developer guide on streaming citations describes the mechanism: the model inserts numbered references like [1] and [2] in the text, and those numbers map to the search results it retrieved. Answers in the Perplexity app also carry numbered source markers, so you can see which page sits behind each statement. A citation in this design is not decoration; it points at a page the answer drew on.

Two fetchers feed this. PerplexityBot crawls and indexes pages ahead of time. Perplexity-User is different: according to the same crawler page, when users ask a question it might visit a web page to help provide an accurate answer and include a link to the page in its response, and because a user requested the fetch it generally ignores robots.txt rules.

The practical lesson is that citations are attached at the level of a sentence or a claim. A page gets cited when it contains the specific statement the answer needs. Long pages that circle a topic without stating anything precise give the model nothing to point a number at.

Open your robots.txt and look for any rule that disallows PerplexityBot, or a blanket User-agent: * block that catches it. Perplexity notes that each robots.txt setting works independently and that it may take up to 24 hours for its systems to reflect changes, so do not judge a fix the same afternoon.

Then check the layer most people forget: your web application firewall or bot protection. Perplexity's crawler page includes setup steps for Cloudflare and AWS WAF and recommends rules that match both the user-agent string and the published IP ranges, set to allow. It publishes those ranges as JSON files and advises using the most current version, ideally refreshed automatically. Matching on both the name and the IP is also how you tell the real crawler from impostors using its user agent.

Finally, read your server logs. Successful requests from PerplexityBot to your important URLs are the only proof that access works. If you only see challenges, redirects to a login, or error codes, Perplexity is not reading those pages no matter what robots.txt says.

Note that Perplexity separates search from training differently from OpenAI. OpenAI's crawler documentation splits ChatGPT search (OAI-SearchBot) from model training (GPTBot). Perplexity states that PerplexityBot is not used for foundation model training at all, so allowing it is purely a search-visibility decision.

Perplexity users tend to ask specific, research-style questions: "does X integrate with Y", "how much does X cost for a team of ten", "what is the difference between A and B for small clinics". Broad category pages compete with every large publisher on the web. Narrow questions are where a smaller site can be the single best source.

Build your list from three places. First, the questions your sales and support teams answer every week. Second, the follow-up questions Perplexity itself suggests under an answer in your category, which show how people refine their research. Third, the sources Perplexity already cites for your topic: if every answer leans on the same forum thread or comparison page, the question it answers is clearly in demand, and a better first-party answer has room to compete.

Then give each important question its own page or its own clearly headed section. One page that answers one question precisely is easier to cite than one page that mentions twenty questions in passing.

Open with the answer in one sentence that would still make sense if quoted alone. Follow it with the specifics a researcher would want to verify: numbers, dates, versions, prices, limits and conditions. Put comparisons in an HTML table with clear headers rather than an image, so every cell is readable text.

Back your own claims with sources. The GEO study by Aggarwal and colleagues, which tested content changes against generative engines including Perplexity, reported that such changes can boost visibility by up to 40% in generative engine responses, and that effectiveness varies by domain. That variation is a reason to test on your own questions instead of copying a template.

Write for the sentence-level citation described above. Short paragraphs that each make one claim, headings that match the questions people ask, and definitions stated plainly near the top of a section all give Perplexity a clean place to attach a number. Remove anything that hides the answer: autoplay pop-ups over the text, content that only appears after a click, or key facts available only in a downloadable PDF.

Many Perplexity questions are time-sensitive, and its search stack tracks when pages were published and updated. Perplexity's date filter documentation for its Search API describes filtering results by original publication date and by when content was last modified or updated, as separate signals. Its Search API overview describes returning real-time, ranked web results. You do not need to know exactly how the consumer product weighs dates to draw the sensible conclusion: a page whose dates are clear and accurate is easier to treat as current.

Show a visible "last updated" date on pages where freshness matters, and mirror it in your structured data with dateModified, which schema.org defines as the date on which the CreativeWork was most recently modified. Change the date only when you change the content. Bumping dates without real edits erodes trust with readers and does nothing for the answer.

Set a review rhythm for pages that carry prices, feature lists, statistics or compatibility details. Those are exactly the facts Perplexity users ask about, and an outdated number is the fastest way to lose a citation to a competitor who kept theirs current.

Your own site is only one of the pages Perplexity can cite about you. Run your target questions and list every domain in the numbered sources: review sites, directories, documentation hubs, community discussions, comparison articles. Those domains are the places your brand needs to be described accurately.

Work through them in order of how often they appear. Complete and correct your profiles, make sure your product is in the comparison tables that exist, answer questions in the communities where your buyers ask them, and keep every description consistent with your own site. When several independent sources state the same facts about you, an answer that draws on any of them describes you correctly. This is cross-source corroboration, and on Perplexity you can watch it happen in the source list.

Ask Perplexity your target questions and open the sources for each answer. Record whether your domain appears, which page is cited, and which competitors are named. Repeat on a fixed schedule, because answers change as the index updates. In your analytics, look for referral visits from Perplexity, and in your server logs, keep an eye on PerplexityBot and Perplexity-User requests to see which pages are being fetched.

To do this across engines at once, start with the free scan at AEO Grader: no signup, about a minute, an AEO score with a four-pillar breakdown and an action plan with code snippets. Its per-engine view is an estimate based on one grounded Gemini check. The $10 one-time report asks your buyer questions to Perplexity, ChatGPT, Claude and Gemini live and shows each engine's actual answer with its sources, plus your share of voice against competitors. Monthly tracking at $9/month repeats that report every month and emails you what changed per engine. No tool can guarantee a citation; what it can do is show you which questions you are winning and which you are not. For the Google side of the same work, see how to show up in Google AI Overviews, and the glossary entry on Perplexity covers the basics.