Perplexity sends two different visitors to your site. PerplexityBot crawls and indexes pages so Perplexity can surface and link them in its answers. Perplexity-User fetches a page live when someone's question needs it, and Perplexity says it "generally ignores robots.txt rules" because a person triggered the visit. To be cited, PerplexityBot has to be allowed. The rest depends less on Perplexity-specific tricks than most guides claim, and more on whether your site already earns organic traffic.
PerplexityBot vs Perplexity-User
Perplexity's bot documentation describes them this way:
| PerplexityBot | Perplexity-User | |
|---|---|---|
| Job | "Designed to surface and link websites in search results on Perplexity" | "When users ask Perplexity a question, it might visit a web page to help provide an accurate answer" |
| robots.txt | Respected. Perplexity recommends allowing it. | "Generally ignores robots.txt rules", since a user initiated the request |
| Training | Not used to crawl content for AI foundation models | Not used for crawling or collecting training data |
| If you block it | Your pages drop out of the index Perplexity cites from | robots.txt alone won't stop it; a CDN or firewall rule can |
Two consequences. First, a robots.txt that disallows PerplexityBot is the one setting that reliably removes you from Perplexity's citations, and the 2024 "block AI crawlers" templates often included it alongside training bots. Neither Perplexity bot is a training crawler, so blocking PerplexityBot buys no training opt-out.
Second, edge-level bot rules can reach the fetcher that robots.txt can't. For new domains onboarding to Cloudflare from 15 September 2026, "Agent" bots, meaning bots acting in real time on a person's behalf, are blocked by default on ad-monetized pages (Cloudflare). Perplexity-User fits that description. Check your CDN settings and your logs for both user agents.
PerplexityBot reads the HTML your server returns. If your content only appears after client-side JavaScript runs, check the raw response with view-source (not the browser inspector, which shows the page after scripts ran).
Why Perplexity is the easiest engine to measure
Because it cites sources on essentially every answer by design. ChatGPT, by contrast, attached a citation to 6.8% of US desktop answers in May 2026, according to Similarweb's 2026 Generative AI Landscape report, up from about 1.3% a year earlier. That average includes writing and coding prompts, but it still means many ChatGPT test runs return no sources at all, while a Perplexity run almost always returns a list you can check your domain against.
Perplexity also has an official API, Sonar, that returns the cited URLs with each answer. A short script can run the same questions every week and log the sources. API answers are not guaranteed to match perplexity.ai, so keep that data labeled as API data. citelity samples Perplexity this way and marks it with an "official API" badge.
Is Perplexity less dependent on Google rankings?
No, and this is the most repeated Perplexity myth. The reasoning goes: Perplexity runs its own crawler, so your Google performance doesn't matter. The first half is true. The conclusion is not.
Ahrefs measured how closely a site's organic traffic tracks how often AI engines mention it (June 2025 data). Perplexity had the strongest relationship of the engines tested:
Its own crawler decides whether Perplexity can find you. Whether it picks you depends heavily on the same authority that earns organic traffic. Keep funding your SEO.
What Perplexity tends to cite
Three engine-specific data points, each with its limits:
- Older reference pages are tolerated. In Seer Interactive's 2026 study of 47,097 citations, 65% of the pages Perplexity cited had been updated within the past year, the lowest of the three engines measured (ChatGPT 73%, Gemini 78%). Seer describes Perplexity as more tolerant of older reference material. A well-maintained evergreen page has a better chance here than elsewhere.
- Community sources are prominent. In Profound's dataset of 680 million citations (August 2024 to June 2025), Reddit was Perplexity's most-cited domain at 6.6% of its citations. That data predates several model changes, so treat it as a tendency, not a current ranking.
- Attribution helped in the one controlled test. The GEO paper (Aggarwal et al., KDD 2024) tested page edits on Perplexity.ai and reported visibility gains of up to 37%, with citing sources, adding quotations and adding statistics among the strongest methods. The test is from 2023–24 and Perplexity has changed since.
The general page work (answer first, self-contained sections, named sources) is the same as for every engine and is in what is AEO.
What AI referral traffic is worth
Perplexity is one of the few engines that sends a clear referrer, so you can see its clicks in analytics. Expect them to be few. Across 13,770 domains, Conductor found AI referrals made up 1.08% of visits, and ChatGPT sent 87.4% of that. The quality numbers that circulate describe AI search traffic in general, not Perplexity: Semrush estimated the average AI-search visitor is worth 4.4× an organic visitor based on conversion rate, across about 500 digital-marketing topics (July 2025), and SE Ranking measured AI visitors spending 67.7% more time on site.
Perplexity is no longer the only place to see AI citation data, either. Bing Webmaster Tools reports citations for Copilot, and Search Console now shows impressions for Google's AI features. Perplexity's referrers remain the most direct view of clicks from Perplexity itself.
Measuring your Perplexity citation rate
- Confirm
PerplexityBotrequests in your server logs over the last 30 days. No hits means an access problem, and nothing else matters until it's fixed. - Pick 10–20 questions your readers actually ask, phrased conversationally.
- Run each several times over a few days (or weekly through the Sonar API) and record whether your domain appears and which sentence was quoted.
- Compare against the same record after any change. How to track AI citations explains how many runs make a rate trustworthy.
How can I stop Perplexity-User from fetching my pages?
Does Perplexity train AI models on my content?
Sources
- Perplexity — Perplexity crawlers (PerplexityBot, Perplexity-User)
- Cloudflare — Content Independence Day: AI crawler options, 1 Jul 2026
- Search Engine Journal — ChatGPT links out most on travel questions (Similarweb 2026 Generative AI Landscape), 27 Jul 2026
- Ahrefs — Websites with more traffic have more AI mentions, Jun 2025
- Seer Interactive — Content recency's impact on AI visibility in 2026
- Profound — AI platform citation patterns (680M citations, Aug 2024–Jun 2025)
- Aggarwal et al. — GEO: Generative Engine Optimization, KDD 2024 (arXiv 2311.09735)
- Conductor — AEO/GEO benchmarks report
- Semrush — AI search SEO traffic study, 21 Jul 2025
- SE Ranking — AI traffic research study
Updated October 1, 2026: corrected PerplexityBot's role (indexing, not answer-time fetching) and added Perplexity-User; added Ahrefs' correlation figures; corrected the Semrush conversion figure's scope; re-attributed the ChatGPT 6.8% figure to Similarweb; removed an invented example answer.