Perplexity SEO: how to get your site cited by Perplexity
Perplexity shows its sources, which makes it the most measurable AI engine. Which bot to allow, why firewalls silently block it and what earns the citation.
Perplexity is the AI search engine that shows its work. Every answer lists the pages it drew on, numbered and linked. For a business, that makes it the most measurable AI platform: you can see exactly which page was cited, for which question, and whether it was yours.
Getting cited comes down to three things. Perplexity’s crawler has to be able to reach you. Your page has to be one of the best available answers to a specific question. And it has to state that answer plainly enough to quote.
Two bots, two jobs
Perplexity documents two user agents (Perplexity docs), and they behave differently:
| PerplexityBot | Perplexity-User | |
|---|---|---|
| Purpose | “Designed to surface and link websites in search results” | “Supports user actions within Perplexity”, fetching a page because a user asked |
| Follows robots.txt | Yes | “Generally ignores robots.txt rules”, because a person requested the fetch |
| Used to train AI models | No, per Perplexity | No, per Perplexity |
Perplexity’s recommendation is direct: allow PerplexityBot in robots.txt if
you want to appear in its results. Changes “may take up to 24 hours” to be
reflected.
User-agent: PerplexityBot
Allow: /The silent failure: your firewall
This is the problem we find most often, and nobody notices it because nothing looks broken. Bot protection on a CDN or host blocks automated traffic it does not recognize. Your robots.txt says “welcome”, your firewall says “no”, and Perplexity simply never sees your pages.
Perplexity publishes the IP ranges its bots use as JSON files and tells site owners to “always use the most current IP ranges” when configuring a firewall. Ask your host or developer to:
- Check server or CDN logs for requests from
PerplexityBotand their response codes. A run of 403 errors is the tell. - Allow Perplexity’s published IP ranges in the bot-protection rules.
- Do the same for OpenAI’s
OAI-SearchBot, which has the same problem.
What earns the citation
Perplexity has not published a ranking formula, and we will not pretend to know one. What we can say is how answer engines like it work in general: they retrieve candidate pages for the question, then pick the passages that answer it most directly and credibly. That rewards pages that make the answer easy to find and easy to trust.
Answer the question in the first lines. If the question is “how long does an AI visibility audit take?”, the best page says “about two weeks” near the top, not after four paragraphs of introduction.
One page, one question. A focused page on “commercial roof coating cost” beats a general services page that mentions coatings in passing.
Be specific and checkable. Figures, timeframes, named methods and dated facts give the engine something to quote. Where a fact is not yours, link the source.
Use tables for comparisons. “X vs Y” questions are answered far more cleanly from a table than from paragraphs.
Keep pages current, honestly. An accurate “last updated” date helps. A fake one, bumped without real changes, is the kind of thing that erodes trust when noticed.
Our CITE framework turns these into a four-step test you can run on any page.
How to measure it
Because Perplexity links its sources, measurement is unusually direct:
- Referral traffic. Visits from Perplexity arrive with
perplexity.aias the referrer, so they show up as their own source in Google Analytics. - Citation checks. Ask Perplexity the questions your buyers ask and note which pages it cites. If a competitor is cited, open their page and compare it with yours for that exact question.
- Repeat over time. Answers change with new content and new questions, so record results monthly rather than once.
Our guide to measuring AI visibility covers building a prompt set and handling the variation between runs.
Questions people ask
Does blocking PerplexityBot keep my content out of Perplexity?
It stops your pages from being surfaced in Perplexity’s search results.
Perplexity-User may still fetch a page when a user specifically asks about
it, because Perplexity treats that as a user action that generally does not
follow robots.txt.
Does Perplexity use my content to train its models? Perplexity states that neither of its bots crawls content for training AI foundation models.
How quickly will changes show up? Robots.txt changes can take up to 24 hours to apply, per Perplexity. Content changes show up once the page is re-crawled, which varies.
Is Perplexity worth the effort if most of my traffic comes from Google? The work that earns Perplexity citations, clear answers, specific facts and open access, is the same work that helps Google’s AI features and ChatGPT. You are not optimizing for one platform, you are making your pages quotable.