How to get featured in Perplexity answers: what the documentation actually says

The short answer

How to get featured in Perplexity answers is a question with a well-documented half and an undocumented half, and most advice on it pretends the second half does not exist.

The documented half is the plumbing. Perplexity publishes exactly what its crawlers are called, what each one does, the user-agent strings, and the IP ranges to verify them against. That part is precise, checkable and worth acting on this week.

The undocumented half is everything about selection. Perplexity’s bot documentation contains no statement about how it chooses, ranks or weights sources. No citation algorithm, no schema preference, no freshness window, no author-credential bonus. The confident percentages circulating about all four are not Perplexity’s, as far as we can establish.

That absence is not a dead end. It tells you the honest answer: there is no Perplexity trick, so the work is the same work that makes you citable everywhere, plus two specific crawler facts that genuinely are Perplexity-specific.

Key takeaways

• Perplexity runs two crawlers with different rules. PerplexityBot indexes for search results. Perplexity-User fetches a page because a person asked about you.
• Blocking one does not block the other. Perplexity’s documentation states that because a user requested the fetch, Perplexity-User “generally ignores robots.txt rules”. Teams who “blocked Perplexity” have often removed themselves from the index while remaining fetchable.
• PerplexityBot is not a training crawler. The documentation says it surfaces and links websites in search results and is not used for training AI models, which changes the calculus for anyone blocking it on principle.
• Verify by IP range, not by user-agent string. Perplexity publishes both ranges as JSON. Anyone can spoof a user-agent string.
• Nothing published tells you how to win a citation, so treat any specific figure about Perplexity’s ranking weights with suspicion, including the ones this page used to carry.

A note on where this comes from. We run organic for more than 250 enterprises and track more than 10 million prompts across every major engine. The crawler facts are Perplexity’s, quoted and dated. Everything about what to do next is our judgement, shaped by the accounts we run. We label it as judgement throughout, because the alternative is inventing numbers.

Book a growth audit if you want to know whether Perplexity is citing you today before you change anything.

What is PerplexityBot, and what does Perplexity actually document?

Perplexity is an answer engine that searches the live web and shows its sources. For a brand, the practical question is what it crawls and how it picks.

On crawling, the documentation is genuinely good. Perplexity names two agents and describes each one.

PerplexityBot, in Perplexity’s words, is “designed to surface and link websites in search results on Perplexity”, and not used for training AI models. Perplexity publishes its user-agent string in full, and its IP ranges as JSON at a fixed URL.

Perplexity-User supports user actions inside Perplexity by visiting web pages to help answer questions. It is not used for web crawling or training. The user-agent string and IP ranges appear the same way.

And then the sentence that matters most, quoted verbatim: “Since a user requested the fetch, this fetcher generally ignores robots.txt rules.”

On selection, there is nothing. The same documentation contains no information about how Perplexity chooses, ranks or weights sources for citation. We looked for a publisher guidelines document as well. Perplexity’s publishers program page returned HTTP 403 to our check on 29 September 2026, so we cannot say what it contains, only that we could not read it. We are not claiming no such document exists. We are saying that a claim nobody can link is not a claim you can use.

The two-crawler mistake that costs people citations

This is the most useful thing on the page, and it follows directly from the quotation above.

A team decides it does not want its content used by AI. Somebody adds a robots.txt rule blocking Perplexity. Everyone moves on.

What actually happened:

• PerplexityBot stopped crawling, so the site drifts out of whatever index Perplexity maintains for search results. Citations fall.
• Perplexity-User carried on, because the documentation says it generally ignores robots.txt when a user requested the fetch. It still fetches pages when somebody asks about the brand.

So the block achieved the opposite of the intention on both counts, and Pepper finds it in about one client audit in five. It reduced the brand’s presence in answers, and it did not stop the fetching that people usually object to.

The reverse case matters too. If you want citations, PerplexityBot has to be able to reach you, and Perplexity’s own description makes the trade explicit: it surfaces and links websites in search results, and it is not used for training models. If your objection to AI crawlers is about training, that objection does not apply to this one on Perplexity’s own account.

Where this falls short: this is Perplexity’s description of its own behaviour. Pepper has not independently audited whether the bots behave as documented, and we would treat “generally ignores” as the operative hedge rather than a guarantee in either direction.

Verify the bot, do not trust the string

A user-agent string is a header. Anyone can send it. Perplexity publishes IP ranges precisely so you do not have to take the string on trust.

1. Pull the published IP ranges. Perplexity provides them as JSON at fixed URLs, one per agent. Treat these as the source of truth.
2. Grep your server logs for both user agents. Confirm whether you are seeing PerplexityBot, Perplexity-User, both or neither. Neither is itself a finding.
3. Check the requesting IPs against the published ranges. Requests carrying the string from outside the ranges are not Perplexity, and should not inform any decision you make.
4. Read your robots.txt as it actually is, not as you remember it. A blanket AI-crawler rule added a year ago is the most common cause of a site disappearing from answers.
5. Check your firewall and CDN separately. Bot-management rules block crawlers that robots.txt permits, and they are the second most common cause. This is the step teams skip.

What this does not tell you. Whether a crawl turns into a citation. That is the undocumented half, and no amount of log analysis answers it.

What we think follows, and it is judgement rather than documentation

Since Perplexity publishes nothing about selection, there are two honest positions. One is to say nothing. The other is to say what we do and label it clearly. We are taking the second, and everything in this section is our opinion.

• Treat Perplexity work as generic AI search work. Perplexity documents no engine-specific ranking factor, so the sensible default is that general quotability carries over here. Self-contained passages, clear answers, real expertise, and coverage on third-party sources.
• Be reachable and be fast. This is the one place the documentation does bear on outcomes, indirectly. A crawler that cannot fetch or render your page cannot cite it, which is why what a crawler actually receives matters more than how the page looks.
• Write passages that survive extraction. An answer engine quotes a passage, not a page, and the passage craft guide is the specific discipline.
• Do not build a Perplexity-only programme. If a tactic only makes sense for one engine, and that engine documents no preference for it, you are optimising for a rumour.
• Measure rather than assume. Our measurement stack covers how to tell whether Perplexity cites you at all, which is the only Perplexity-specific number we would trust: your own.

What would change our view. A published statement from Perplexity about source selection. Until then, ask anyone quoting a percentage for its citation weights to produce the link. The answer usually ends the conversation.

Our methodology: how we judge an engine-specific claim

Weights sum to 100. Publication and cross-engine value carry 60 between them. That is Pepper’s judgement rather than a fact, and it is the judgement this page rests on. We publish our weightings the same way we publish [how we rank GEO agencies](https://www.pepper.inc/blog/geo-agency-ranking-methodology/), so you can argue with them precisely.

What would change these weights. A published statement from Perplexity about source selection would reorder the whole list, because the first criterion would suddenly have something to point at. Until then it does the most work.

The claims we removed, and why

We are naming these rather than quietly deleting them, because the page carried them under a Pepper byline and readers may have acted on them.

Seven rows, covering thirteen figures.We have kept nothing that we could not open at source, and the page is shorter on numbers and longer on things you can check as a result.

How to get featured in Perplexity answers, at a glance

What this costs

• The crawler audit: about twenty minutes, free, and it needs nothing you do not already have.
• Fixing a robots.txt or firewall block: an hour, usually less, and it is the highest-return item on this page.
• Everything else is general AI search work, costed on our generative engine optimization cost benchmarks rather than as a Perplexity project.
• What you should not pay for: a Perplexity optimisation package built on ranking factors nobody has published. Ask for the link before you sign.

Where Pepper fits

We are an agentic organic growth engine, and our honest position on this engine is that we do not have private knowledge of it.

• An organic growth partner. A growth team is attached to your account, a senior strategist plus always-on agents, accountable for the number rather than for an engine-specific tactic list.
• Pepper’s GEO platform. Genuinely self-serve and included rather than billed as a separate licence. You log in and set up a workspace. Define your brand profile, competitors and personas, then connect GA4 and Search Console, and manage your own themes, prompts and analytics across six engines including Perplexity, ChatGPT, Gemini, Claude and Google AI Overviews.
• The Agent Atlas. Your team builds, versions and runs its own agents, and keeps them.

What we can actually offer on Perplexity is measurement rather than secrets. Tracking whether it names you and whether it cites your domain, across a frozen prompt set, is a real number about your brand. A published ranking weight is not available to us or to anyone else. For what the combination produces, the Acceldata case study records 6X organic traffic and top-three keywords rising from 85 to more than 300, mostly through B2B SaaS programmes.

Where it falls short: we cannot tell you how to win a Perplexity citation specifically, because nobody outside the company can, and we published a page that implied we could. That is the least comfortable sentence in this article. We are also not a technical practice, so if the finding is a CDN rule blocking a crawler, we will identify it and your engineering team will fix it.

How to choose what to do about Perplexity

I will write this in the first person, because a recommendation nobody will defend is worth nothing.

Start by finding out whether Perplexity cites you at all, before changing anything. It is the first thing Pepper checks on a qualification call. Most teams asking this question have never checked, and the answer reorders the work entirely. If Perplexity cites you regularly, the job is protecting that. If it never does, the job is usually reachability or third-party coverage rather than anything engine-specific.

So my order is: check the crawlers, check your own citations, then do general work. Here is the scorecard I would use on any proposed Perplexity tactic, on a 100 point scale.

Five questions I would ask about any Perplexity advice, in this order.

1. “Where is the link?” A stronger answer opens a Perplexity page. A weaker one names a document nobody can produce, which is what the previous version of this page did.
2. “Is PerplexityBot able to reach us?” Check robots.txt, the firewall and the CDN. This one question resolves more cases than the rest combined.
3. “Did we block Perplexity at some point?” Ask the person who administers robots.txt, not the marketing team. They often say yes, and nobody remembered.
4. “Does Perplexity cite us now, and for which questions?” Run your own prompt set. It is the only Perplexity-specific number available to you.
5. “Would we still do this if Perplexity did not exist?” If yes, it is real work. If no, ask question one again.

Red flags, each one something a supplier actually says. A percentage for Perplexity’s citation weights, offered without a link. “Perplexity favours Reddit” as a strategy rather than an observation. A schema package sold as a Perplexity citation lever, when Google states plainly that structured data is not required for generative AI search. A guarantee of citations, which nobody can promise. And an engine-specific retainer for an engine that publishes no ranking guidance.

Before changing anything, measure. Write 25 real commercial questions your buyers would type. Run each of those 25 prompts three times in Perplexity specifically, because engines are probabilistic and one run is an anecdote. Then give any fix 90 days before you judge it. Four worked examples for a mid-market B2B software company:

• “best data observability platform for enterprise”
• “how do I monitor data quality across Snowflake and Databricks”
• “what should I look for in a data observability vendor”
• “who competes with Monte Carlo in data observability”

Read the sources Perplexity cites, because it shows them. That is this engine’s genuine advantage for a marketer: you can see exactly which pages won, which is a better guide than any published weighting would be. Checking whether your brand appears in AI search walks through the running part.

And the answer that loses Pepper the sale: if PerplexityBot can reach you and Perplexity already cites you, you do not need a Perplexity programme and you do not need us for one. Spend the money on earned media, because the sources Perplexity shows you are mostly not going to be your own site.

It comes down to one principle: do the documented thing, and refuse the undocumented number. The previous version of this page did the opposite, ours included in the count of pages that got this wrong, and it is the reason this one is shorter on statistics.

What nobody should promise you

• A citation weight for schema, freshness or author credentials. Perplexity publishes none, and we could not open the document those figures point to.
• A guaranteed citation. Engines are probabilistic and nobody controls their selection.
• That blocking robots.txt stops Perplexity fetching your pages. Its own documentation says Perplexity-User generally ignores robots.txt when a user requested the fetch.
• A Perplexity-specific tactic that does not help elsewhere. If it only works here and here documents nothing, it is a guess with a price attached.
• Market share or query volume figures without a named source. The previous version of this page carried several, and we could not verify any of them.

Frequently asked questions

How do I get featured in Perplexity answers?
Make sure PerplexityBot can reach you, then do general AI search work: self-contained answers, real expertise and third-party coverage. The company publishes nothing about how it selects sources, so any specific ranking tactic a supplier offers you is unsourced.

What is the difference between PerplexityBot and Perplexity-User?
PerplexityBot surfaces and links websites in Perplexity search results. Perplexity-User visits pages to answer one specific user’s question. Perplexity states that neither agent is used for training models, and that Perplexity-User generally ignores robots.txt rules.

Does blocking Perplexity in robots.txt work?
Partly, and probably not as intended. It stops PerplexityBot crawling, which removes you from the index feeding answers. Its documentation says Perplexity-User generally ignores robots.txt when a user requested the fetch, so it may still fetch your pages.

Does Perplexity use my content to train models?
Perplexity’s documentation states that PerplexityBot is not used for training AI models, and that Perplexity-User is not used for web crawling or training either. That is Perplexity describing its own behaviour, which we have not independently audited.

Does schema markup help me get cited by Perplexity?
There is no published evidence that it does. Perplexity documents nothing about schema, and Google states structured data is not required for generative AI search. Keep schema for rich results rather than as a citation tactic.

How do I verify traffic is really from Perplexity?
Check the requesting IP against the ranges Perplexity publishes as JSON, one file per agent. A user-agent string is just a header and anyone can send it, so a string alone is not verification of anything.

Does Perplexity favour Reddit and forums?
It often cites them, and you can see that yourself because Perplexity shows its sources. Whether that reflects a documented preference is unknown, and we could find no published origin for the specific percentages circulating about it.

How do I know if Perplexity is citing my brand?
Run a frozen set of 25 real buyer questions three times each in Perplexity. Record whether it names you, and whether it cites a page from your domain. That is the only Perplexity-specific measurement available to you, and it is free.

Where to go next

If you want the plain explanation of what Perplexity is before optimising for it, our Perplexity explainer covers the basics.

If the finding is that crawlers cannot reach your content, the three tiers of hidden content covers what a crawler actually receives. If they can reach you and you are still not quoted, how to write a passage that survives being quoted is the next constraint, and the GEO measurement stack tells you which of the two you have.

Sources and further reading

• Perplexity, “Bots” documentation, read at source on 29 September 2026. All Perplexity quotations on this page are verbatim from it. PerplexityBot: “designed to surface and link websites in search results on Perplexity”, not used for training AI models, with a published user-agent string and IP ranges at perplexity.com/perplexitybot.json. Perplexity-User: supports user actions by visiting web pages to answer questions, not used for web crawling or training, with its own published string and ranges. On robots.txt: “Since a user requested the fetch, this fetcher generally ignores robots.txt rules.” The page carries no publication or update date, and it contains no information about how Perplexity selects or ranks sources.
• Perplexity publishers program page. Returned HTTP 403 to our check on 29 September 2026. We could not read it and therefore make no claim about its contents in either direction.
• Google Search Central, “Optimizing your website for generative AI features on Google Search”, last updated 10 July 2026. Used once, for the position that structured data is not required for generative AI search, which bears on one of the removed claims.

A note on the statistics that are not here. Thirteen figures from the previous version were removed. Several circulate widely on third-party marketing blogs attributing them to Perplexity publisher guidelines from February 2026. We could not find that document on any Perplexity property we could reach, and none of the blogs quoting it links an original. We are not asserting the document does not exist. We are declining to quote a source we cannot open, which is the standard we would apply to anyone else.

Further reading on Pepper, each checked live on 29 September 2026:

• What Perplexity AI is, the explainer that sits under this page.
• The three tiers of retrievability, if reachability is the constraint.
• How to structure content for answer engines, if quotability is.
• The GEO measurement stack, to tell the two apart.