Short answer
Perplexity is an AI answer engine that searches the web for each question and writes a reply with numbered source links, so optimizing for it means being easy to find, read and cite. Allow PerplexityBot, keep key facts in plain HTML text, answer questions directly with specifics, and earn mentions on other sites. No trick guarantees a citation, and the same basics work across AI search engines.
How does Perplexity find and cite sources?
Perplexity treats each question as a fresh search. It looks up pages, pulls out the passages that seem to answer the question, and a language model writes a reply that points to numbered sources. It always shows those sources, so a citation is visible and clickable. Perplexity does not publish how it ranks or chooses them, so treat any “Perplexity ranking formula” as a guess.
Because it retrieves pages at question time, a good new page can be cited soon after it is crawled, without waiting for a model to be retrained. That makes crawlability and freshness matter more than they do for answers drawn from a model's memory.
Which Perplexity crawlers should I know?
PerplexityBotcrawls the web to build the search index Perplexity uses to surface and link sites. Perplexity says it is for search results and not for training its models.Perplexity-Userfetches a page when a person's question calls for it. Perplexity says that, because a person asked, this fetcher generally ignores robots.txt rules.
If you want to be cited, do not block PerplexityBot. Check two layers: your robots.txt, and your host's bot protection, which can turn crawlers away before they read it. Which AI crawlers to allow or block has a table of the main names and a sample file.
What helps a page get cited?
- Be crawlable and readable
Pages should load without a login and show their main text in the HTML. Many AI crawlers do not run scripts, so text that appears only after JavaScript may be invisible to them.
- Lead with the answer
Put a question as a heading, then a direct answer of about 40 to 75 words, then the detail. Passages that stand alone are easier to quote.
- Add specifics
The 2024 Princeton and IIT Delhi paper found that adding statistics and cited sources raised visibility in generated answers by up to roughly a third, while keyword stuffing did worst of the tactics tested.
- Show who wrote it and when
Name the author, show a last-updated date and refresh stale facts. Engines that retrieve live can favor current, clearly attributed pages.
- Be present beyond your site
In many categories, answers lean on review sites, directories, forums and video, so honest mentions there matter. Check which sites Perplexity cites for your questions.
- Rank in ordinary search too
Ahrefs found about 76% of pages cited in Google's AI Overviews ranked in Google's top ten. That figure is for Google, and Perplexity runs its own index, so the overlap may differ, but sound SEO remains the base.
What does not help?
- llms.txt. Google does not use it and no major AI answer engine has confirmed reading it; see whether llms.txt works.
- Schema markup as a citation trick. Ahrefs' May 2026 study of 1,885 pages found no meaningful lift in AI citations after adding JSON-LD, although those pages already had many citations. Schema still has other uses.
- Keyword stuffing and hidden text. Both are manipulation, and the first was the weakest tactic in the research above.
- A promised Perplexity position. Nobody can sell one, because answers change from run to run.
How do I check whether Perplexity cites me?
- Write 10 to 20 questions your customers would ask, plus a few that name your brand.
- Ask each one in Perplexity on different days and open the sources list.
- Note whether your domain is a source and whether the text names your brand.
- Look for visits from perplexity.ai in your analytics.
Arcline can automate the same loop on Premium: it stores each answer with the assistant, date and sources, and shows whether it mentions and links to you. Which assistants it asks depends on the AI search services connected at the time. Its can-they-read-it check fetches your pages the way the crawlers of Perplexity, ChatGPT, Claude and Google do; see the features page.
Does this apply to other AI search engines?
Mostly yes. ChatGPT search, Claude with web search, Gemini and Google AI Mode each run their own crawlers and indexes, but they reward the same things: access, clear text, specifics and trust. Google says the same SEO fundamentals apply to its AI features. A Perplexity-only strategy is rarely worth it. Start with how to show up in ChatGPT answers and how AI assistants choose sources.