Perplexity shows its sources. Make sure one of them is you.
Perplexity is built around live search and puts its sources front and center. That makes it one of the clearest places to see who's beating you, and why.
Search first, then answer
Perplexity was built as an answer engine from day one. Search isn't a feature it reaches for occasionally; it's the core of the product. Perplexity looks for relevant pages, reads them, and writes a summary with citations attached. Paid users can pick from several underlying models, but retrieval does the heavy lifting on which sources make the cut.
Its own index and its own crawlers
Perplexity runs its own search index, fed by its crawler, PerplexityBot. In 2025 it opened that index to developers through a Search API. A second agent, Perplexity-User, fetches pages live when a person's question needs them.
PerplexityBotcrawls to surface and link sites in results. Perplexity recommends allowing it and says it doesn't crawl content for AI model training.Perplexity-Useracts for a person, so Perplexity says it generally ignores robots.txt.- Perplexity publishes IP ranges for both agents, which helps when you allowlist them in a firewall.
- Robots.txt changes can take up to 24 hours to register, according to Perplexity.
How citations display
Citations are the point of Perplexity. Answers carry inline citations next to the claims they support, and the full source list is a tap away. Suggested follow-up questions keep the thread going, often with fresh searches and fresh sources. Because each claim needs evidence, pages that state facts clearly and specifically are easy to cite, and vague pages are easy to pass over. For a primer on how engines choose evidence, see grounding.
5 ways to win in Perplexity.
- 01
Let PerplexityBot through
Allow it in robots.txt and allowlist Perplexity's published IP ranges in your firewall. A blocked crawler means a thin or missing presence in the index Perplexity searches.
- 02
Make every claim citable
Perplexity attaches sources to individual statements. Write specific, self-contained sentences with names, numbers and dates you can stand behind, so one line can serve as the citation.
- 03
Keep pages fresh and dated
Perplexity searches live, so recency is visible to it. Show publish and update dates, refresh pages when facts change, and retire stale versions that still get crawled.
- 04
Earn the sources it already trusts
Look at the domains Perplexity cites for your prompts. If a review site, forum or publication keeps appearing without your brand in it, that's where to earn coverage next.
- 05
Answer the follow-up
Perplexity suggests related questions after each answer. Cover those next questions on your pages, under clear headings, so you're still in the thread three questions later.
Does Perplexity use my content to train AI models?
Perplexity says neither PerplexityBot nor Perplexity-User crawls content for AI model training. PerplexityBot exists to surface and link websites in Perplexity's search results. Allowing it means opting into being found and cited, not into training. Check Perplexity's crawler documentation for the current wording before you set policy.
Can I block Perplexity-User?
You can add a rule, but Perplexity says Perplexity-User generally ignores robots.txt because a person requested the fetch. If you truly need to stop it, firewall rules based on Perplexity's published IP ranges are the stronger tool. Think twice first: those fetches happen because someone asked about your page, which is usually a page you want read.
Why does Perplexity cite my competitor's page instead of mine?
Common reasons: their page answers the exact question more directly, it's easier to crawl, it's fresher, or a third-party page names them and not you. Compare the cited page with yours line by line: the opening answer, the specifics, the date. Elite Answers' source analysis shows which pages win each prompt over time.
How fast do robots.txt changes reach Perplexity?
Perplexity says its systems can take up to 24 hours to reflect robots.txt changes. Recrawling and re-indexing take their own time after that, so expect a lag before citations move. Make the change, confirm in your server logs that PerplexityBot is fetching pages successfully, and judge the results over the following weeks.
Somewhere, right now, an AI is recommending your client's competitor.
Elite Answers is invite-only. Request access through Elite Digital and the team will set you up.