# Perplexity SEO: how to get your site cited by Perplexity

> Perplexity SEO is getting your pages cited in Perplexity's answers. The crawler that decides it, 201 links from 10 answers, and what to change.

Canonical: https://picked.so/perplexity-seo
Last updated: 2026-09-29

Perplexity SEO is getting your pages cited, and your brand named, in Perplexity's answers. Perplexity searches the web for every answer and links many sources, about 20 per answer in the ones we recorded, so the job is to let PerplexityBot crawl you and to be on the pages it reads.

- **20.1** links per answer, against ChatGPT's 8.0 for the same questions
- **25%** of links went to a brand the answer named
- **0** links to Reddit, in 201

10 answers for LaunchOnIt through Perplexity's Sonar API with web search, US, September 27 and 28, 2026.

> Track the real app, not the API, where you can.

Our Perplexity figures come from its Sonar API with web search, not the app, and this page says so, as Picked does next to the number. A tracker that calls an API answer Perplexity without saying so reports something your buyers never see.

## How does Perplexity choose which sites to cite?

Perplexity is an answer engine: every question goes to a web search first, and the answer is written from the pages it finds, with numbered citations. Unlike ChatGPT or Claude, it doesn't wait to decide whether a question needs the web. It always searches.

That makes Perplexity the most search-like of the AI models, and the one where a page can earn a link fastest. It also means its citations spread wide: in our answers, Perplexity linked far more pages than any other model, and most of them weren't the brands it named.

## Which Perplexity crawlers need to reach my site?

Two, according to [Perplexity's crawler docs](https://docs.perplexity.ai/guides/bots), read September 29, 2026:

| Crawler | What it does | Block it and |
| --- | --- | --- |
| `PerplexityBot` | Finds and links websites in Perplexity's search results. Perplexity says it isn't used to train AI models | Your site doesn't appear in Perplexity's search results |
| `Perplexity-User` | Opens a page when a person's question needs it | Perplexity can't fetch your page for that person; it generally ignores `robots.txt`, since a person asked |

Allow `PerplexityBot`. It's the one that decides whether you can be cited, and since it doesn't collect training data, there's little reason to block it. Perplexity publishes the IP ranges of both, so a firewall can let the real ones in and keep impostors out.

## What does Perplexity cite?

We asked Perplexity 5 buyer questions for LaunchOnIt, a launch platform we run, on September 27 and 28, 2026, US: 10 answers. These came through Perplexity's Sonar API with web search, not the Perplexity app, which can answer differently. Read them as how Perplexity's search picks sources, not as screenshots of the app.

| What we counted | Perplexity |
| --- | ---: |
| Links per answer (ChatGPT, same questions: 8.0) | 20.1 |
| Different sites linked | 80 |
| Links to the site of a brand the answer named | 51 of 201 (25%) |
| Links to pages titled like a list ("best", "top", "alternatives", "vs") | 81 of 201 (40%) |
| Links to Reddit (ChatGPT: 2 of 80) | 0 |

The most-linked site was another launch platform's own "best launch platforms" lists (26 links), then an AI news site (15), Product Hunt (8) and "awesome" lists on GitHub (7). Perplexity linked the brands it named a quarter of the time. The rest went to the pages that compare them.

So for Perplexity, the pages that win citations are the ones that rank and compare: roundups, directories, "X vs Y" posts, curated GitHub lists. A brand that's on those pages gets named; a brand that writes one of them gets cited. Ten answers on one product is a small sample, so check it against your own.

**Sources**: LaunchOnIt's answers on Perplexity, 10 answers, 201 links from 80 sites.

| Site | Page | Times cited |
| --- | --- | ---: |
| smollaunch.com | 16 Best Product Hunt Alternatives in 2026 (Free + Paid) | 26 |
| aiweekly.co | AI News: Artificial Intelligence Stories, Ranked | 15 |
| producthunt.com | Best of Product Hunt Competitors & Alternatives (2026) | 8 |
| ai-tldr.dev | Where to Promote Your App in 2026 | 7 |
| github.com | rushout09/directory-submission-sites: 250+ free Startup ... | 7 |
| launchon.it |  | Not cited |

## How do I get my site cited by Perplexity, step by step?

No one can make Perplexity cite you. These steps give it reasons to.

1. **Allow `PerplexityBot`.** In `robots.txt`, your CDN's AI bot settings and your firewall.
2. **Write the comparison page for your category.** A dated "best X for Y" list or "X vs Y" page that names every option honestly, you included, with prices and who each is for. Perplexity links these more than anything else.
3. **Put the answer first.** Perplexity quotes sentences. A first line that answers the question with a fact (a price, a number, a date) is the line that gets a citation number next to it.
4. **Get onto the other lists.** For your brand to be named, it has to appear on the roundups and directories Perplexity reads. Ask Perplexity your buyers' questions and open every source: that's the list to work through.
5. **Keep pages current.** Show the date you last updated a page with prices or rankings, and update it. An out-of-date price gives an answer a reason to quote someone else.
6. **Keep doing SEO.** Perplexity's results come from a search index; pages that are indexed, linked and fast get found.

## How do I know if Perplexity is citing me?

**Ask it.** Your buyers' questions, on several days. Note which of the numbered sources are yours.

**Read your logs.** `PerplexityBot` hits mean you're being indexed. `Perplexity-User` hits mean people are asking questions that make Perplexity open your pages right now.

**Track it.** An [AI visibility tool](https://picked.so/) asks the same prompts on a schedule and records which pages each answer links. Picked asks Perplexity every day, through the Sonar API with web search, and labels it as such next to the number; it also counts `PerplexityBot` and `Perplexity-User` on your site. Where the real product's answer can be read, as for [ChatGPT](https://picked.so/chatgpt-seo), Picked reads that instead. A tracker that calls an API answer "Perplexity" without saying so is reporting something your buyers never see.

## When Perplexity SEO won't pay off

**Your buyers don't use Perplexity.** Ask your customers which assistants they use before spending a quarter on this one.

**You're chasing a citation, not a customer.** Twenty links per answer means yours can be number 17. Being cited is a start; being named is what sends buyers.

## Frequently asked questions

### Is Perplexity good for SEO?

It's a useful place to be cited, because it searches the web for every answer and links many sources. It's also a quick way to research a topic: the sources it lists show which pages a search engine trusts for a question. It doesn't replace Google SEO; its results come from a search index built the same way.

### Does Perplexity use Google?

Perplexity runs its own crawler, PerplexityBot, to build the index its answers draw on, and says that crawler decides whether sites appear in its results. Ranking on Google helps indirectly, since pages that rank are usually linked and indexed widely, but it isn't the same list.

### How do I rank higher in Perplexity?

There's no fixed rank: each answer links the sources it used. To be linked more often, allow PerplexityBot, publish the comparison or list page for your category with dated facts, answer the question in the first sentence, and get your brand onto the roundups Perplexity reads.

### Does Perplexity train AI on my website?

Perplexity says PerplexityBot isn't used to crawl content for AI foundation models; it's for search results. Perplexity-User fetches pages when a person asks a question. If you want to stay out of AI training, other companies' training crawlers, such as GPTBot and ClaudeBot, are separate settings.

### Why does Perplexity cite my competitor instead of me?

Usually because the pages Perplexity reads for that question mention your competitor and not you, or your competitor wrote the comparison page itself. Ask Perplexity the question, open each source, and note which ones leave you out. Those pages, and the comparison you haven't written yet, are the fix.
