SEO

Perplexity SEO: How Perplexity Finds, Picks and Cites Sources, From Its Own Docs

Perplexity SEO explained from Perplexity's own help pages and crawler docs: how it searches the web, what PerplexityBot and Perplexity-User do, what its Government, Academic and Trusted source labels mean, and what a site owner can control. Plus which popular claims Perplexity has never confirmed.

Alexander ChuaCo-Founder, Growigami November 13, 2026 9 min read

“Perplexity SEO” covers two jobs. The first is getting your pages read and cited when Perplexity answers a question. The second is using Perplexity to do SEO research, which is what Perplexity’s own SEO use-case page is about. This guide is mostly about the first job, with a short section on the second.

When we read the top Google results for “perplexity seo” on 13 November 2026, two were agency pages and one was a three-year-old Reddit thread. The agency guides repeated claims such as “Perplexity cites 3 to 4 sources per answer” and “Perplexity keeps pre-set lists of authoritative domains” without a link to anything Perplexity has published. Below, every statement about how Perplexity works links to Perplexity’s own help center or crawler documentation, read on 13 November 2026. Where Perplexity has said nothing, we say so.

How does Perplexity find and pick its sources?

Perplexity describes its answer process in four steps in its help article How does Perplexity work? (last updated May 1, 2026):

  1. It interprets the question, using language models to work out what you are asking.
  2. It searches the web in real time, gathering information from “authoritative sources like articles, websites, and journals.”
  3. It summarizes the most relevant points into one answer.
  4. It cites its sources. “Each answer includes numbered citations linking to the original sources.”

Its What is Perplexity? page adds that content “is sourced from the web in real-time as you ask your questions” and that responses are supported by citations from “reputable news organizations, academic publications, and established content sources.” The Search product page says every search “pulls from our live web index” and that Perplexity runs more than 20 models, which it post-trains “to search harder, find better evidence, and answer more clearly.”

Deeper modes read more pages. Perplexity says Deep Research “performs dozens of searches, reads hundreds of sources” and writes a report in 2 to 4 minutes. A page that is not picked in a quick answer can still be read in a research run on the same topic.

What Perplexity does not publish: a ranking formula, a number of citations per answer, or a list of preferred domains. Treat any guide that states these as fact as a guess unless it links to a Perplexity source.

What are PerplexityBot and Perplexity-User?

Perplexity documents two user agents on its crawler page. They do different jobs and follow robots.txt differently.

User agentWhat Perplexity says it doesFollows robots.txt?Used to train AI models?
PerplexityBot”Designed to surface and link websites in search results on Perplexity”Yes. Perplexity recommends allowing it “to ensure your site appears in search results”No. “It is not used to crawl content for AI foundation models”
Perplexity-UserVisits a page when a user’s question needs it, and may link that page in the answer”Generally ignores robots.txt rules,” because a user requested the fetchNo

Three details from the same page matter in practice:

  • Changes take up to 24 hours. Perplexity says it may take up to 24 hours for its systems to reflect a robots.txt change.
  • Firewalls block it more often than robots.txt does. Perplexity publishes IP ranges for PerplexityBot and Perplexity-User and recommends allowing them in your web application firewall, matching each user agent to its own IP list. It notes that Cloudflare’s Bot Fight Mode “can’t be skipped with WAF custom rules,” so it can still challenge Perplexity’s bots.
  • Spoofed bots exist. Matching both the user-agent string and the published IP ranges is how Perplexity suggests telling its real crawler apart from a scraper using its name.

A robots.txt group that keeps your site open to Perplexity’s search crawler looks like this:

User-agent: PerplexityBot
Allow: /
Disallow: /admin/

If you want different rules for each AI company, our free robots.txt generator writes the groups for OpenAI, Anthropic, Perplexity and Google in one file.

What do Perplexity’s source labels mean?

Some Perplexity citations carry a shield icon with one of three labels: Government, Academic or Trusted. Perplexity explains them in Understanding source labels (last updated August 7, 2026). This is the closest thing Perplexity publishes to a statement about which sites it considers reliable.

  • Labels rate whole websites, not single pages. A label describes the domain.
  • The tests are editorial. Perplexity’s examples: “Does the site correct its mistakes? Does it say who wrote each piece? Does it keep news separate from advertising and opinion?”
  • Money does not buy a label. “Perplexity’s partnerships, payments, and other business arrangements do not affect a site’s label.”
  • No label is not a penalty. Most domains have none, and Perplexity says that “is not a negative judgment.”

For a B2B company the practical reading is simple. Name an author on every article, publish a corrections or update note when you change a fact, and keep sponsored content clearly marked. These are the same habits Google’s guidance rewards, and they are the habits Perplexity says it checks.

What can a site owner control?

LeverWhat to doSource for why
Crawl accessAllow PerplexityBot in robots.txt; allow its IP ranges in your firewall and CDNPerplexity crawler docs
Pages that answer the questionPut the answer, with its number and source, in the first sentence under each headingPerplexity says it “compiles the most relevant insights” into a summary (help center)
FreshnessKeep facts that change (prices, versions, dates) current and datedPerplexity says it sources content “in real-time” (help center)
Editorial signalsNamed authors, visible corrections, ads kept apart from contentThe source-label tests (help center)
Mentions on other sitesBe cited by the articles, journals and established publications Perplexity readsPerplexity names these source types itself (help center)

Most of this overlaps with what you would do for ChatGPT search. The differences are the crawler names and how each company describes its sources; our ChatGPT SEO guide covers OpenAI’s side from OpenAI’s own documentation, and our glossary entry on GPTBot explains why blocking OpenAI’s training crawler is a separate decision from appearing in ChatGPT answers.

We found these in the top results and could not find them in any Perplexity help page or doc on 13 November 2026:

  • “Perplexity cites 3 to 4 sources per answer.” Perplexity publishes no number. Deep Research reads “hundreds.”
  • “Perplexity uses pre-set lists of authoritative domains by category.” The only published system is the source-label review, which Perplexity says rates sites on editorial tests and does not change because of payments.
  • “Schema markup is essential to be cited.” Perplexity’s docs do not mention structured data. Schema is still useful for Google; it is just not a Perplexity requirement on record.
  • “Perplexity copies Google’s rankings.” Perplexity says it searches a “live web index.” Some third-party studies report overlap between Perplexity citations and Google’s top results, which is what you would expect when both reward the same well-sourced pages.

How do you measure Perplexity citations?

Perplexity offers no equivalent of Google Search Console, so measurement takes three checks:

  1. Referral traffic. In your analytics, filter sessions whose referrer is perplexity.ai. People who read an answer and then type your URL show up as direct traffic, so this number is a floor.
  2. Server logs. Look for the PerplexityBot and Perplexity-User user agents and confirm the IPs match Perplexity’s published lists. If PerplexityBot never appears, check your firewall before you rewrite any content.
  3. Answer checks. Keep a fixed list of 20 to 50 questions your buyers ask. Run them in Perplexity once a month, log whether each answer cites you, which competitors it cites and which URLs. Run each question twice, because answers change. Our guide to AI visibility tools compares the software that automates this.

How do you use Perplexity for SEO research?

This is the other meaning of the keyword. Perplexity’s SEO use-case page suggests prompts for trending topics, competitor content, search intent and content gaps. Three that work well because Perplexity cites what it found:

  • “What questions do [job title]s ask before buying [category] software? List each with the source where it was asked.”
  • “Compare how [competitor A] and [competitor B] explain [topic] on their own sites. Link each page.”
  • “What changed in [topic] in the last 90 days? Only use sources dated in that period.”

Perplexity cannot see your search volumes, rankings or Search Console data, so check its numbers against a keyword tool before you act on them.

Where Growigami fits

About half of B2B software buyers now start their research in AI chatbots, according to G2’s March 2026 survey. Growigami gets B2B companies named there: AI agents write the pages ChatGPT and Perplexity cite from what your team knows, and a CMO or CRO ties the work to pipeline. If you want to see which buyer questions already cite your competitors, our AI search and SEO service starts with that list.

Frequently asked

What is Perplexity in SEO?

Two things. Perplexity is an AI search engine that answers questions by searching the web and citing its sources with numbered links, so 'Perplexity SEO' usually means getting your pages read and cited there. It is also a research tool SEO teams use for topic, competitor and search intent research; Perplexity publishes a use-case page for that second meaning.

How does Perplexity choose its sources?

Perplexity's help center says it interprets the question, searches the web in real time, gathers information from 'authoritative sources like articles, websites, and journals', summarizes the most relevant points and adds numbered citations. It does not publish a ranking formula. It does publish the crawler it uses for search results (PerplexityBot) and a source-label system that rates whole websites on tests such as whether they correct mistakes and say who wrote each piece.

Does Perplexity use Google's results?

Perplexity does not say so in its help center or crawler documentation. Its search product page says every search 'pulls from our live web index', and its crawler docs describe PerplexityBot as the bot that surfaces and links websites in Perplexity's search results. Claims that Perplexity copies Google or Bing rankings come from third parties, not from Perplexity.

Should I block PerplexityBot?

Only if you do not want to appear in Perplexity's answers. Perplexity's docs say PerplexityBot is designed to surface and link websites in search results and 'is not used to crawl content for AI foundation models.' Blocking it does not protect your content from model training, because that is not what it collects for. It does make your pages ineligible to be found by Perplexity's search crawler.

Is optimizing for Perplexity different from SEO for Google?

Most of the work is the same: pages that crawlers can reach, clear answers, named authors and real sources. Two things are specific to Perplexity: allowing PerplexityBot and Perplexity's published IP ranges through robots.txt and your firewall, and the source labels Perplexity shows next to some citations. Perplexity does not offer a webmaster console, so you measure it through referral traffic and by asking it your buyers' questions.

How do I see if Perplexity cites my site?

Look in your analytics for sessions referred by perplexity.ai, check your server logs for the PerplexityBot and Perplexity-User user agents, and keep a fixed list of buyer questions to run in Perplexity each month while logging which answers cite you. Answers change between runs, so run each question at least twice.

Alexander Chua

Co-Founder, Growigami

Decade running growth at B2B SaaS. Co-founded Growigami to ship the work every week.

Start with one process.

We learn the business and build your first go-to-market process on us this month.