How Perplexity Chooses Its Sources

Perplexity is built around citations: nearly every answer shows numbered sources. That makes it the easiest AI engine to study. It also means being among the sources is the whole game. Here is what Perplexity has published, what we can observe, and what we cannot know.

What Perplexity publishes

Perplexity documents two user agents (Perplexity crawler documentation):

• PerplexityBot is described as the crawler that surfaces and links websites in Perplexity search results. Perplexity states it is not used to train foundation models.
• Perplexity-User fetches a page when a user's question calls for it. Perplexity notes that, because the visit is user-initiated, it may not follow robots.txt in the same way.

Check the page above for current wording, since vendors change these terms.

What is not published

Perplexity does not publish a ranking formula for which pages become sources. Anyone who claims to know the exact weights is guessing. What we can do is observe the sources in real answers and look for patterns. Treat patterns as observations about specific queries on specific days.

What we typically see to look for

Rather than claim a pattern, here is what to check in your own results:

• Which domains repeat. For local service questions, run several variations and list the cited domains. If the same directories, review sites, or forums appear again and again, those are your listing targets.
• Whether the cited page answers the question directly. Open the cited pages. Notice how they are written: a clear heading, a direct answer, specifics.
• Whether competitors' own sites are cited. If so, compare their page to yours for facts, clarity, and depth.
• Freshness. Note the visible dates on cited pages.

Steps to become citable

1. Allow the crawler

Make sure robots.txt does not block PerplexityBot. Some sites block it along with training crawlers by accident. If you choose to block it, accept that it removes one route to being cited.

2. Make pages self-explanatory

A retrieval system pulls passages. A passage that makes sense on its own, such as "A standard roof inspection in Orlando takes 30 to 60 minutes and covers..." (use your real numbers), can be quoted. A passage that depends on a header image or a slogan cannot.

3. Be specific about place and service

Say which cities and neighborhoods you serve, which services you perform, and any limits. Vague pages match fewer queries.

4. Get listed where the answers point

Because Perplexity shows its sources, you can see exactly which third-party pages it uses for your category. Make sure you are accurately present on them.

5. Keep pages fast and accessible

Fetchers time out. A page that takes many seconds to respond or needs a login may be skipped.

Reading Perplexity results fairly

• It varies. Ask twice. Our paid report runs each question twice per engine and reports "consistent", "sometimes", or "no".
• Mode matters. The consumer site has different modes and settings. SearchSignal uses Perplexity's developer API with web search on (or an API reseller when no direct key is set), so results can differ from perplexity.ai. We state this on the methodology page.
• Mention and citation are different. Your name can appear without a link to your site, or your site can be a source without your name in the text. We count and show both.

A short test you can run today

1. Write three questions a customer would ask about your service in your city.
2. Ask each in Perplexity, logged out.
3. Copy the list of sources for each answer into a table.
4. Count which domains appear more than once.
5. Check whether you are listed, accurately, on each of those domains.

That table is a to-do list. Work through it over a few weeks, then repeat the test with the same questions.

Automating it

SearchSignal runs this test across Perplexity and four other engines, shows the sources each one cited, and (in the full report) ranks the most-cited domains across all answers so you can see where to be listed. Start with the free check.

Common questions

Does Perplexity use Google or Bing?

We have not seen Perplexity publish a complete list of its search sources, so we do not claim one. It does document its own crawler, PerplexityBot.

Why is my competitor cited and not me?

Open their cited page and compare it with yours. Look at clarity, specifics, freshness, and whether other sites also mention them. Then look at where else they are listed.

Do paid Perplexity plans change results?

Settings and models can differ by plan and mode. That is one reason API-based measurement can differ from what a given user sees.

Can I ask Perplexity to include me?

No. You can make your pages reachable and clear, and you can fix your listings. Inclusion is the system's decision.

Published 2026-10-30. We cite vendor documentation where we state facts about it; vendors change products often, so check the linked pages for current details. This guide describes what you can influence and does not promise any result.