How to get recommended by Perplexity: a step-by-step method

Perplexity says it answers from live web content gathered when a question is asked, and its crawler is meant to surface and link websites. It publishes no formula for recommendations. The practical method is to test your current answers, confirm that PerplexityBot can reach your site, write passages that answer questions directly with named sources, and re-test on a fixed schedule.

The essentials

• Log a fixed set of questions a real buyer would ask, and run them in Perplexity before you change anything.
• Allow PerplexityBot in robots.txt, then check your server logs for successful requests.
• Put a direct answer with a named source in the first two or three sentences of each page.
• Re-test the same question set every week or every month, and judge trends rather than single answers.

What do you need before you start?

Work through the steps in order. Each one depends on the one before it.

• Live, crawlable pages: Perplexity's help center says content is sourced from the web in real time as you ask. Your page has to be live, crawlable and clear when a question comes in.
• Known limits: According to AuthorityTech's analysis, Perplexity does not disclose exact weights, candidate counts, thresholds or a fixed source-selection formula. Some guides, such as Stackmatix, publish precise percentage weights for ranking factors. Perplexity has not confirmed those numbers, so don't plan around them.
• A measurable goal: Perplexity names you for the questions your buyers ask, it links to your site, and the wording it uses about you is accurate.

How do you test what Perplexity says about you today?

• What to do: Write a set of questions a real buyer would ask, such as "best [service] for [city]" or "[product type] for small teams". Run each one in Perplexity, copy the answer and the cited domains, and log the date. Keep the wording identical for every re-test.
• How to know it's right: Your log shows, for each question, whether you are named, whether your domain is cited, and which competitors appear instead. That is your baseline.
• Typical mistake: Deciding from one answer. The same logic applies as in checking ChatGPT with more than one prompt.

Can PerplexityBot reach your site?

"PerplexityBot is designed to surface and link websites in search results on Perplexity." — Perplexity, in its crawler documentation

• What to do: Open your robots.txt. Perplexity recommends allowing PerplexityBot in that file and permitting requests from its published IP ranges. Check that your CDN or firewall doesn't block those ranges.
• Why it matters: The same documentation says PerplexityBot is not used to crawl content for AI foundation models. Allowing it is a search-visibility decision, not a training one.
• How to know it's right: Your server logs show PerplexityBot requests returning successful responses, and your robots.txt rules match what you intended. The AI crawler checker shows whether robots.txt blocks Perplexity, ChatGPT or Claude.
• Typical mistake: Believing robots.txt controls every visit. The documentation says Perplexity-User, the agent that fetches a page when a user asks a question, generally ignores robots.txt rules. Treat robots.txt as a signal for the crawler, not a lock on user-triggered visits.

Does the first paragraph answer the question?

• What to do: On each page that should be cited, put a direct answer to its main question in the first two or three sentences. Support it with specifics: the number, the date and the named source. Keep sentences short.
• Why it matters: One analysis from Fokal says Perplexity chooses sources for their pertinence to the specific claim being made, not for overall domain authority. A passage with one clear claim is easier to match to a question than a long page that mixes several. Citable's guide to Perplexity sourcing covers the same logic.
• What the evidence shows: Kevin Indig studied 7,000 citations and found that the top 10% of cited pages had higher sentence counts, word counts and Flesch scores. That is a correlation, not proof of cause. Use it as a reason to write clearly, not as a target.
• Typical mistake: Padding the page to reach a length, or rewording the same question in several headings. That reads as text written for search engines, and it doesn't help a person decide.

Are your facts and sources easy to check?

• What to do: Make sure your name, address, services, prices, opening hours and contact details match everywhere they appear. Update outdated pages first, because Perplexity reads live content.
• Your own sources: When you cite a statistic, name its source and year, and link to the original.
• Trust labels: Perplexity marks some citations with one of three labels: Government, Academic or Trusted. The same commentary warns they are trust context, not a score that explains why a page was chosen.
• Where analyses disagree: eSEOspace says Perplexity favours .edu and .gov domains for technical queries, which conflicts with the claim-level view above. Build on clear, checkable pages and don't bet on either theory.
• On llms.txt: Citable's llms.txt explainer covers what the file does and whether it helps, so treat it as optional.
• Typical mistake: Shipping a schema or markup change and assuming it moved Perplexity's answers. Check your log before you call it a success.

How do you get mentioned beyond your own site?

• What to do: Get your business described accurately on sites your buyers already read: industry publications, local directories, partner sites and forums where your work comes up.
• Why try it: Fokal's vendor analysis claims Perplexity tends to name brands that appear consistently across forums, publications and industry sites, not just on their own domain. Perplexity hasn't confirmed this, so treat it as a working hypothesis that complements your own pages.
• How to know it's right: Your log shows mentions of your brand on domains that aren't yours, and you can explain why each one is there.
• Typical mistake: Paid placements that pass as editorial coverage, or reviews you wrote yourself. Those aren't independent mentions.

How often should you re-test, and what if you miss?

• What to do: Repeat your baseline question set every week or every month. Log the date, the answer, the cited domains and any change since the last run. Change one thing at a time, so you can link a result to an action.
• Why misses happen: Being retrieved is not the same as being cited. AuthorityTech's summary of research on Perplexity's Sonar model says it visited about ten relevant pages per query but cited only three or four.
• Why nobody can promise a result: In 2024, Perplexity's CEO admitted to Fortune that he didn't have the answer on how its AI decides which products to recommend.
• Typical mistake: Judging the whole method from one bad answer, or celebrating one good one. Look at trends across the set over several weeks.

Next steps

• Write your buyer questions today and run them in Perplexity. Save the answers and cited domains as your baseline.
• Open your robots.txt and your CDN or firewall rules. Confirm PerplexityBot is allowed and its requests return successful responses in your logs.
• Rewrite the first two or three sentences of your five most important pages so each gives a direct answer with a named source.
• List three sites your buyers already read, and plan one accurate mention on each.
• If you want the weekly testing handled for you, Citable does it this way: it measures whether ChatGPT, Gemini, Perplexity, Claude, Grok and Google AI cite or recommend your site, and finds out why competitors are picked instead. Plans start at $24/month per site.

Frequently asked questions

Can I block PerplexityBot and still appear in Perplexity answers?

Perplexity describes PerplexityBot as the crawler designed to surface and link websites in its results, and recommends allowing it. Blocking it removes the route Perplexity built for that purpose. Perplexity-User, which fetches pages when a user asks, follows different rules.

Can robots.txt stop Perplexity from visiting my site?

Not reliably. Perplexity's documentation says Perplexity-User, the fetcher triggered by a user's request, generally ignores robots.txt rules. For PerplexityBot, the documentation treats robots.txt as the place to allow it.

How do I check whether Perplexity cites my website for my main customer questions?

Write a fixed set of buyer questions and run them in Perplexity. Log whether you are named, whether your domain is cited and which competitors appear. Repeat the same set on a schedule and compare trends.

Why does Perplexity recommend competitors when my site has better content?

One analysis says Perplexity picks sources for their pertinence to each specific claim, so a competitor's passage may match the question more directly. Research summarised by AuthorityTech also says Perplexity visited about ten pages per query and cited only three or four.

Sources

• Perplexity crawler documentation
• What is Perplexity? (Perplexity Help Center)
• AI Platforms Like ChatGPT and Perplexity Decide What to Cite (AuthorityTech)
• How Perplexity Decides Which Sources to Cite (AuthorityTech)
• Perplexity AI Optimization Strategy: Citation Guide (Stackmatix)
• How Perplexity Decides Which Sources to Show in 2026
• How Does Perplexity Choose Sources? (Fokal)
• How Perplexity Chooses the Sources It Quotes (eSEOspace)
• Perplexity AI ranking factors: A guide for SEOs (Keyword.com)
• Fortune, 2024: Perplexity CEO on product recommendations