# How to get cited by Perplexity: sources, crawlers and pages

> In short: Perplexity searches the web for nearly every question and lists the pages it used next to the answer. To be one of them, allow PerplexityBot in robots.txt, rank for the specific searches it runs, and put a direct, specific answer near the top of the page. Because its sources are visible, Perplexity is also the quickest place to see which third-party pages decide your category.

Guide from Keats, published 2026-11-13, updated 2026-11-13. Canonical: https://getkeats.com/guides/perplexity-citations

Of the major assistants, Perplexity is the most transparent about where its answers come from. It's built as an [answer engine](https://getkeats.com/glossary/answer-engine) that searches first: ask almost anything and the answer arrives with numbered sources, each one a page you can open. That makes it the easiest assistant to study, and the most direct to influence.

## How a Perplexity answer is put together

Perplexity reads the question, searches the web, picks a set of pages, and writes the answer from them with a numbered [citation](https://getkeats.com/glossary/citation) after the claims each page supports. The sources are listed with the answer, so for any question you can see exactly which pages it drew on.

Two consequences follow. Its answers move with the web more quickly than a model's memory does, because they're rebuilt from search each time. And the list of cited pages for your buying questions is, very nearly, the list of pages you need to be on.

## Step 1: let PerplexityBot in

Perplexity documents two user agents. **PerplexityBot** "is designed to surface and link websites in search results on Perplexity", and Perplexity says it isn't used to crawl content for training foundation models. **Perplexity-User** fetches a page when a user's question needs it; because a user asked, Perplexity says it "generally ignores robots.txt rules".

So if you want to be cited, allow PerplexityBot in robots.txt, and make sure your CDN or firewall isn't blocking it either. Perplexity notes that robots.txt changes may take up to 24 hours to apply. Our guide to [AI crawlers and robots.txt](https://getkeats.com/guides/ai-crawlers-robots-txt) has the full list for every assistant.

## Step 2: study the sources for your questions

Ask Perplexity each of your buying questions several times and record every cited page. Group them by domain. In most categories you'll see three kinds:

- **Publications and roundups**: articles comparing options for a use case.
- **Communities**: Reddit threads, forums, Q&A sites.
- **Vendors' own pages**: product, pricing, comparison and help pages.

Mark which pages name you, which name rivals and not you, and which get your facts wrong. This is the work list, and Perplexity gives it to you more directly than any other assistant.

## Step 3: make your pages easy to cite

Perplexity attaches citations to specific claims. A page gets cited when it contains a sentence that supports a claim the answer is making. That favours pages that:

- **State the answer near the top**, in plain words that match the question.
- **Contain specific, checkable facts**: numbers, prices, dates, named features, comparisons.
- **Are clearly about one thing**, so a search for that thing finds them.
- **Show their evidence**: data, methods, sources. The 2023 paper that named generative engine optimization tested rewrites on a benchmark of queries: its three top methods, citing sources, adding quotations and adding statistics, improved visibility in the generated answers by 30 to 40%. Tested on Perplexity itself, adding quotations did best, at 22%, while keyword stuffing did 10% worse than leaving the page alone.

Our guide to [writing content for AI search](https://getkeats.com/guides/answer-first-content) has the page structure.

## Step 4: earn the third-party citations

For buying questions, most of Perplexity's sources aren't vendors. Getting onto them is outreach, not optimization: a fact sheet for a roundup's author, a correction where they have you wrong, honest participation in the threads buyers read. Our guide to [getting on the lists AI cites](https://getkeats.com/guides/third-party-placements) covers each kind.

## What's different about Perplexity

| | Perplexity | ChatGPT |
|---|---|---|
| Searches the web | On nearly every question | When the question needs current information |
| Shows sources | Always, numbered, beside the answer | Usually, when it searched |
| Search crawler | PerplexityBot | OAI-SearchBot |
| How quickly answers follow the web | Fast: rebuilt from search each time | Fast for searched answers, slow for memory |

The practical upshot: Perplexity is where a change to a page shows up soonest, and where you can confirm a page is being used by seeing it in the sources.

## Measuring it

Track your buying questions on Perplexity separately from the others; averaging assistants hides where you're strong and weak. Count the share of answers that name you and the share that cite one of your pages, over several answers per question per week. Clicks from Perplexity arrive with perplexity.ai as the referrer, which a GA4 channel can group with other assistants.

For the mechanics of how the other assistants pick sources, see [how ChatGPT search picks sources](https://getkeats.com/guides/chatgpt-search-sources).


## Your agent does this

Ask your agent: "Which domains does Perplexity cite most on our buying questions, and which of them name a rival but not us?" It calls `read_by_engine` for Perplexity and `list_citations`, and lists the pages worth a correction or an outreach note.

## Sources

- [Perplexity: Perplexity crawlers](https://docs.perplexity.ai/guides/bots), read 2026-11-13
- [Aggarwal et al., GEO: Generative Engine Optimization (arXiv 2311.09735)](https://arxiv.org/abs/2311.09735), read 2026-11-13