Skip to content
husky hamster

How to Secure Perplexity Citations for Your Brand

Rafał Moszkowcow (Chomsky)CEO & Co-Founder
Andreea PaluszakHead of Outreach
Published , updated 4 min read

The short answer

How do you get cited in Perplexity?

Make your pages reachable by Perplexity's crawler, make them worth quoting, and keep your facts consistent wherever they appear. Perplexity's documentation (n.d.) says PerplexityBot surfaces and links websites in Perplexity's search results and recommends allowing it in robots.txt and permitting requests from its published IP ranges. Beyond access, Perplexity publishes no ranking guide, so the rest is third-party evidence: a Princeton study found that adding statistics, quotations and citations raised source visibility in generative engines, up to 37% on Perplexity itself, in a research setting (Aggarwal et al., 2024). No tactic guarantees a citation.

What are Perplexity citations and brand mentions?

A brand mention is your name appearing in an answer, with or without a link. A citation is a page attached as a source for the answer. A mention without a citation, or a citation without a mention, is common, so track both. Citations are not ranked results: the same question can produce different sources on different runs.

What does Perplexity say about its crawlers?

Perplexity’s documentation (n.d.) describes two user agents.

Agent What it does Robots.txt
PerplexityBot Designed to surface and link websites in search results on Perplexity. Not used to crawl content for AI foundation models Perplexity recommends allowing it
Perplexity-User Supports user actions. When a user asks a question it might visit a page and link to it. Not used for crawling or for training foundation models Generally ignores robots.txt, because a user requested the fetch

Each has a published list of IP addresses on perplexity.com. Perplexity says each robots.txt setting works independently and may take up to 24 hours to take effect.

Check your firewall. If you use a web application firewall, Perplexity’s documentation suggests allowing requests that match both its user-agent strings and its published IP ranges, and updating the ranges regularly (Perplexity, n.d.). Then check your server logs for PerplexityBot requests that return errors instead of normal responses.

How do you make a page worth citing?

Google’s guide is about Google, but its central advice is a fair test for any engine: unique, non-commodity content with a distinct point of view and first-hand evidence (Google, 2026a). Specific ways to do it:

  • Answer one question per section, clearly. This helps readers first. Google says chunking is not required for its systems, and we found no verified evidence that it matters for Perplexity.
  • State claims with scope and evidence. The Princeton GEO paper found that adding citations, quotations and statistics raised visibility on its benchmark, up to 40% overall and up to 37% on Perplexity, with effects that varied by domain (Aggarwal et al., 2024). Treat it as support for evidence-rich writing, not a promise.
  • Show who wrote it and when it was last reviewed. Keep the dates accurate.
  • Make key facts readable in the page text, not only in images or scripts.
  • Write the page yourself. Do not publish claims, numbers or “results” you cannot show. If a figure is an example, label it as one.

How do you keep facts consistent?

If your site says one thing and a review site or directory says another, you give any engine conflicting information to choose from. List the facts buyers ask about (price, plan limits, free trial, support hours, certifications, data location), then compare them across your site, review sites, directories and any other profile. The checker below compares what you type, ignoring case, spaces and punctuation. Its example rows are placeholders.

Brand fact consistency checker

Type the same fact as it appears in each place. Rows compare after ignoring case, spaces and punctuation. The example rows are placeholders. Nothing leaves your browser.

FactYour siteReview siteDirectoryOtherResult
Consistent rows
-
Rows with a mismatch
-
Rows with fewer than two values
-

Text version of the checker. For each fact, enter the value as it appears on your site, a review site, a directory and one other place. If two or more values differ once case, spaces and punctuation are ignored, the row is a mismatch. Rows with fewer than two values cannot be compared. Fix mismatches at the source, starting with the profile you control.

What do third-party mentions have to do with it?

Perplexity’s documentation does not say how it weighs sources, and we found no primary data on it (Perplexity, n.d.). The best public evidence across engines is correlational: Ahrefs’ study of 75,000 brands (Linehan, 2025b) found branded web mentions correlate 0.656 to 0.709 with mentions in ChatGPT, AI Mode and AI Overviews, and Ahrefs states that correlation is not causation.

Maciej Chmurkowski (Enterprise Product Owner at Senuto) reports that community sources such as Reddit and forums dominate ChatGPT citations in his studies, and describes Perplexity as strongly oriented to shopping and e-commerce (Chmurkowski, 2026). These are one practitioner’s observations.

We pursue natural mentions only where they fit the topic and the publication, and we do not manufacture them. Take part in communities honestly, keep review-site profiles accurate, and ask customers for genuine reviews. Our brand mentions service follows the same rule (a link to our own service).

How do you measure it?

Run a fixed prompt set across several days or weeks, record mentions and citations, and report the rate with its uncertainty (see our sampling calculator). Chmurkowski reports that a brand’s first place held in only about 30% of answers when the same prompts were repeated every half hour, so a single check can mislead (Chmurkowski, 2026). In analytics, filter referral sessions from perplexity.ai and compare them with your sampling results.

What mistakes prevent citations?

  • Blocking PerplexityBot at the firewall without noticing (Perplexity, n.d.).
  • Measuring one answer on one day.
  • Treating visibility like a search ranking position.
  • Publishing promotional claims with no evidence, and inconsistent product facts across profiles.
  • Assuming schema markup guarantees a citation. We found no source that says it does.

What does a 30-day plan look like?

Our suggestion, not a measured result.

  • Days 1 to 7: write the prompt set from real buyer questions and run a first sample. Check robots.txt, the firewall and the server logs for PerplexityBot (Perplexity, n.d.).
  • Days 8 to 14: run the fact checker and fix mismatches at the source.
  • Days 15 to 24: improve the pages that answer your highest-value questions, adding evidence and dates.
  • Days 25 to 30: sample again and compare presence by question, not by position.

Key takeaways

  • Perplexity has two agents. PerplexityBot surfaces and links sites in its search results and is not used to crawl content for AI foundation models (Perplexity, n.d.). Perplexity-User supports individual user requests and generally ignores robots.txt, because a person asked for the fetch.
  • If you block PerplexityBot, you reduce your chance of appearing (Perplexity, n.d.). Perplexity recommends allowing it and permitting its published IP ranges, and says robots.txt changes can take up to 24 hours to take effect.
  • A firewall can block it without you knowing. Perplexity's documentation suggests combining user-agent and IP checks in your firewall rules (Perplexity, n.d.).
  • The Princeton GEO paper found that adding citations, quotations and statistics raised visibility on a benchmark by up to 40%, and by up to 37% on Perplexity (Aggarwal et al., 2024). It is a research setting, and effects varied by domain.
  • Keep prices, features and certifications identical across your site, review sites and directories.
  • Measure by repeated sampling. One answer proves nothing, because AI answers vary from run to run.

References

/ faq

Questions people ask

Does Perplexity use robots.txt?

PerplexityBot respects it, and Perplexity recommends allowing it to appear in results (Perplexity, n.d.). Perplexity-User generally ignores it, because a user asked for the fetch.

Is PerplexityBot used to train AI models?

Perplexity's documentation says PerplexityBot is not used to crawl content for AI foundation models, and that Perplexity-User is not used for crawling or for training. (Perplexity, n.d.)

How long do robots.txt changes take?

Perplexity says up to 24 hours.

Does adding statistics and quotations help?

In the Princeton GEO paper it did, in benchmark tests, up to 37% on Perplexity (Aggarwal et al., 2024). Real-world results will vary, and nothing guarantees a citation.

Tell us which pages need to win

Thirty minutes with Andreea. Bring your top three commercial pages, and we'll tell you what's standing between them and page one, whether or not you hire us.

Book a strategy meeting →