How-toChatGPT, Gemini & co.
Perplexity SEO: How Perplexity Picks Its Sources
by the AEO GEO Labs team5 min read
Perplexity SEO is the work of getting your pages cited in Perplexity's answers. Perplexity searches its own web index for every question, ranks passages rather than whole pages, and attaches numbered citations to the sources it used. So the job is simple to state: let PerplexityBot crawl you, rank for the question and its follow-ups, and hold the clearest passage that answers it.
This is part of our LLM SEO guide, which covers every major AI engine. Here we focus on what Perplexity itself has documented, as of October 2026.
How Perplexity picks its sources
Perplexity picks sources by searching its own index, reranking the results at passage level, and citing the passages its model actually uses. Its help center describes the flow in four steps: understand the question, search the web, summarize the most relevant material, and cite sources with numbered links.
The technical detail comes from Perplexity's write-up on its Search API, which runs on the same infrastructure. As of that write-up:
- Perplexity's index tracks over 200 billion unique URLs.
- Retrieval combines lexical (keyword) and semantic (meaning) search, then merges the results.
- Ranking happens in multiple stages, and sections and spans of documents are ranked as units in their own right.
That last point matters most for Perplexity SEO. Perplexity is choosing paragraphs, not domains. A sharp, specific paragraph on a small site can be cited over a vague page on a big one.
Perplexity's two crawlers, and why it matters
Perplexity runs two user agents, and only one of them builds the index. Its crawler documentation describes them:
| User agent | What it does | Respects robots.txt |
|---|---|---|
| PerplexityBot | Crawls to "surface and link websites in search results on Perplexity" | Yes |
| Perplexity-User | Visits a page when a user's question needs it | Generally no, per Perplexity's docs |
If you block PerplexityBot, your pages are unlikely to enter Perplexity's index, and you lose the steady path to citations. Perplexity-User may still fetch a page on demand, but that isn't a strategy. Check your robots.txt for both names with our AI Crawler Checker. Perplexity also publishes IP ranges for each bot, so a firewall rule can allow them without opening the door to impostors.
Perplexity SEO, step by step
The steps below follow the order Perplexity's pipeline runs in: crawl, retrieve, rank, cite.
- Allow PerplexityBot in robots.txt and in any bot-protection rules on your CDN. A firewall that blocks unknown bots can block it even when robots.txt allows it.
- Serve core content as HTML. Don't rely on JavaScript to render the answer text.
- Rank in traditional search for the question. Perplexity's citations overlap with classic search results more than other assistants do. Ahrefs' 2025 study found 28.6% of Perplexity's cited links were in Google's top 10 for the same prompt, against about 8% for ChatGPT.
- Cover the follow-up questions. Perplexity's Pro Search researches across multiple sources for a single question. A page that answers the likely next three questions has more ways to be retrieved.
- Write answer-first passages. Start each section with a 40 to 60 word answer that names its subject. That's the span the reranker scores.
- Use exact numbers and dates. Specific claims are easier to attach a citation to than adjectives like "fast" or "affordable".
- Keep pages current. Perplexity searches the web live for each question, so a current fact beats a stale one. Update facts when they change, and show the update date.
- Earn mentions on sources Perplexity reads. Reviews, forums, and industry publications widen the set of pages that can name you.
Our view: step 3 is underrated. Of the major assistants, Perplexity looks most like a search engine, so ordinary SEO work carries over to it more directly than it does to ChatGPT.
How Perplexity SEO differs from ChatGPT optimization
Perplexity SEO and ChatGPT optimization share the same goal but differ in where the sources come from.
| Perplexity | ChatGPT search | |
|---|---|---|
| Index | Perplexity's own | OpenAI's crawl plus third-party search providers |
| Crawler to allow | PerplexityBot | OAI-SearchBot |
| Citation style | Numbered inline citations | Inline source links and a sources list |
| Overlap with Google top 10 (Ahrefs, 2025) | 28.6% | About 8% |
The tactics in our guide on how to rank in ChatGPT mostly apply here too. The difference is emphasis: with Perplexity, classic rankings and fresh, specific passages do more of the work.
How to track Perplexity citations
Track Perplexity citations by asking a fixed list of 20 to 50 buyer questions each month and logging whether you're cited and who is cited instead. Perplexity shows every source in the answer, so this is quicker than on most engines. Clicks from Perplexity typically appear in analytics as referrals from perplexity.ai.
For the bigger playbook on becoming a source engines trust, see how to earn AI citations.
Frequently asked questions
Does Perplexity use Google or Bing results?
Perplexity describes its own index of over 200 billion URLs in its Search API write-up, built by PerplexityBot. It hasn't published a full list of any outside data sources it may also use. What we can say, based on Ahrefs' 2025 study, is that Perplexity's citations overlap with Google's top 10 more than ChatGPT's, Gemini's or Copilot's do.
Should I block Perplexity-User?
Usually not. Perplexity-User fetches a page when a person's question needs it, so blocking it can stop Perplexity reading your page in that moment. Perplexity's documentation says this fetcher generally ignores robots.txt anyway, because a user triggered the request. If you must restrict it, use a firewall rule with Perplexity's published IP ranges.
Can I pay to appear in Perplexity answers?
Not for organic citations, as of October 2026. Perplexity runs a Publishers' Program that shares revenue with partner publishers when their content is referenced, but that is a payment to publishers, not a way to buy citations. Citations in answers come from retrieval and ranking, which you influence through crawl access and content quality.
Does llms.txt help with Perplexity SEO?
No public evidence says so. As of October 2026, Perplexity hasn't documented using llms.txt to choose citations. Its crawler documentation covers robots.txt and IP ranges, not llms.txt. A file costs little to publish and may help agents, but don't expect it to change your Perplexity citations.
If you want this done for your site, AEO GEO Labs runs GEO, AEO and SEO programs for B2B and SaaS teams. See our services.