Citebase
Guides Pay $197

Perplexity’s own words

How to get cited by Perplexity

Perplexity searches the web each time someone asks a question, and every answer includes citations. For site owners it documents one thing clearly: if you want to be surfaced and linked in its results, allow PerplexityBot and its IP ranges. It does not publish how it chooses which of the pages it found to cite.

What Perplexity documents

Perplexity’s help center describes the product in a few lines. It says it searches the web to give conversational answers, that each response includes citations and links to original sources, and that content is sourced from the web in real time as you ask. It combines several sources into one answer instead of handing back a list of pages. That is from What is Perplexity?

The site-owner detail is on the Perplexity Crawlers page:

PerplexityBot
Designed to surface and link websites in Perplexity search results. Perplexity says it is not used to crawl content for AI foundation models. They recommend allowing it in robots.txt, and allowing their published IP ranges, if you want to appear in results.
Perplexity-User
Fetches a page when a person’s question calls for it, and may link that page in the answer. Perplexity says it is not used for crawling or for training, and that because a user asked for the fetch, it generally ignores robots.txt.
Timing and IP ranges
Each robots.txt setting works independently, and changes can take up to 24 hours to show. Current IP ranges are published as JSON at perplexity.com/perplexitybot.json and perplexity.com/perplexity-user.json. Perplexity says to always use the current lists, because they change.

Nothing on these pages lists ranking factors, a submission form for ordinary sites, or a paid route into answers.

Step 1: make sure PerplexityBot can actually reach you

This is the step most often broken, and it is usually not robots.txt.

  1. Read robots.txt

    Open https://yourdomain/robots.txt. If there is no User-agent: PerplexityBot group, the wildcard group applies. A group of its own is clearer:

    User-agent: PerplexityBot
        Allow: /

  2. Check the firewall

    Perplexity documents allow rules for two common firewalls, which tells you how often this is the problem. For Cloudflare WAF, it describes a custom rule that matches the PerplexityBot or Perplexity-User user agent and the published IP ranges, with the action set to Allow. For AWS WAF, it describes IP sets built from the published ranges, a string match on each user agent, and allow rules placed at a higher priority than any blocking rules. It recommends matching on both user agent and IP, so you are not letting in anything that only claims to be the bot.

  3. Look at the logs

    Search your server or CDN logs for PerplexityBot. If the requests get 403 or a challenge page, the firewall is the reason, whatever robots.txt says. Perplexity notes that WAF changes take time to propagate, so check again a day later.

Blocking PerplexityBot is a choice some sites make on purpose. If you do, know what it does and does not cover. On Perplexity’s own description, it removes you from what PerplexityBot surfaces in search results. It does not stop Perplexity-User, which generally ignores robots.txt. The crawler guide has PerplexityBot next to the OpenAI and Google tokens.

What Perplexity-User visits tell you

Perplexity says Perplexity-User fetches a page when a person’s question calls for it. So a Perplexity-User request in your logs means someone asked a question that led Perplexity to open that page. Perplexity does not describe it this way, but that makes these requests a rough signal of which pages real questions are reaching. Group the requests by URL over a month. Pages that get fetched often are worth keeping especially current. Pages you expected to see and never do may not be matching the questions buyers ask.

Step 2: give it a page worth citing

Once Perplexity can fetch you, the open question is whether your page is the best thing it found for the question. Perplexity does not say how it decides. What follows is observed practice, not Perplexity guidance.

  • Answer the question the buyer asked. If the prompt is “form tool that takes payments for a workshop,” a page that says in its first lines that your form tool takes payments, and what it charges, is matchable. A homepage slogan is not.
  • Keep facts current and dated. Perplexity sources content when the question is asked. An old price on your pricing page will be the price it finds.
  • Make facts easy to lift. Prices, limits, supported platforms, and who the product is for read better in a short list or a table than inside a long paragraph.
  • Link sources for your own claims. If you say something is required by law or published by a vendor, link the source. A research paper on generative engines found that adding cited sources and statistics made pages more visible in answers on its test setup. The GEO guide covers what that paper did and did not show.

Step 3: read the citations you already get

Perplexity shows its sources, which makes it the easiest engine to learn from. Run five buyer questions using the manual check, and for each answer open every cited link. Sort them into a few buckets:

  • Your own pages.
  • Competitor pages.
  • Roundups, reviews, and comparison posts.
  • Forums and community threads.
  • Documentation and reference pages.

If the citations are mostly roundups that leave you out, rewriting your homepage will not change much. The pages that need to be accurate are those roundups. The competitor guide covers how to approach them honestly. If the citations are your own pages but the answer still gets you wrong, the page that was cited is the one to fix.

Step 4: when Perplexity gets you wrong

Perplexity’s help center has a page on reporting incorrect answers. It says to use the flag icon below the answer, or open a support ticket from account settings. Include the URL of the query, a description of the error, and what you expected. Misinformation and outdated information are among the problems it lists.

Fix the source first. If the wrong fact came from your own old page, update it before you report. If it came from someone else’s page, a report may help, but the page is still out there.

If you sell products

In a November 2024 post, Perplexity announced shopping features and a merchant program. The post says product cards are not sponsored, and that the merchant program is free. It says merchants in the program get a better chance of being recommended, payment integration, free API access, and a dashboard of search and shopping trends. At the time, the signup form was aimed at large retailers. That is the announcement as published. Check the current program terms before you plan around it.

Measuring Perplexity visits

Perplexity’s pages above do not document a tracking parameter like the one OpenAI adds for ChatGPT. Site owners commonly see these clicks as referrals from perplexity.ai. That is what people report, not something Perplexity documents.

Google Analytics 4 has an AI Assistant channel in its default channel group. Google’s description names ChatGPT, Gemini, DeepSeek, Copilot, and Grok as examples and says the channel uses a list of AI assistants it does not publish. Perplexity is not named in that text. Look at where your perplexity.ai sessions land before you build a report on the channel.

A short checklist

  1. robots.txt has a PerplexityBot group that allows the pages you want cited.
  2. Your firewall allows PerplexityBot by user agent and published IP range. Logs show 200 responses.
  3. Key pages state what you sell, for whom, and at what price, in current text.
  4. You have saved the citations for five buyer questions and sorted them by type.
  5. Wrong answers are flagged, and the source page behind them is fixed where you control it.

Related: showing up in ChatGPT search, and AI search vs SEO.

Sources