PerplexityBot
Indexes pages so they can be cited in Perplexity's answers. Perplexity states it does not train foundation models.
Reference reviewed by Sona
| Operator | Perplexity |
|---|---|
| Powers | Perplexity's answer engine and citations |
| Purpose | Search & answers |
| User-agent token | PerplexityBot |
| Respects robots.txt | Yes |
PerplexityBot builds the index Perplexity draws on when it answers questions, and pages it indexes can appear as numbered citations beneath answers. Perplexity's own description is that the bot is "designed to surface and link websites in search results on Perplexity."
One claim on Perplexity's crawler documentation deserves attention: neither of its bots trains foundation models. If your concern is training exposure specifically rather than AI visibility generally, Perplexity is a different case from OpenAI, Anthropic, Google, or Meta - there is no training crawler here to opt out of, so the only decision is whether you want to be citable.
That makes PerplexityBot an unusually easy allow for most publishers. Perplexity is citation-forward by design - answers carry visible, numbered source links - so the referral behavior is more like a search engine than a chatbot that absorbs your content and moves on.
Perplexity states PerplexityBot respects robots.txt. Its reputation for compliance has been contested in the past, and its user-triggered sibling Perplexity-User is documented as generally ignoring robots.txt, so keep the two tokens distinct when you write rules and when you read logs.
How PerplexityBot behaves
- Indexes for citation, so it revisits pages that change rather than crawling once.
- Perplexity documents that this bot does not collect content for foundation-model training.
- Impersonation is common enough that Perplexity publishes IP ranges specifically to distinguish genuine traffic.
- Reads its own robots.txt group, separately from Perplexity-User.
Full user-agent string
Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; PerplexityBot/1.0; +https://perplexity.ai/perplexitybot)
Allow PerplexityBot
Perplexity citations are a real referral channel with visible source links, and Perplexity states the content is not used to train foundation models.
User-agent: PerplexityBot Allow: /
Block PerplexityBot
You don't want your content represented or summarized in Perplexity answers at all.
User-agent: PerplexityBot Disallow: /
How to verify PerplexityBot
Published IP range feed
Perplexity publishes PerplexityBot's ranges as JSON, and its current documentation cites the feed on perplexity.com rather than the older perplexity.ai host - worth checking if you automated this a while ago, since a stale feed URL that 404s tends to fail open. Perplexity has publicly flagged traffic impersonating its crawler, so the IP check is the difference between honoring a real indexer and trusting a scraper with a borrowed name.
Check an IP against this botCommonly confused with PerplexityBot
PerplexityBot respects robots.txt; Perplexity-User is documented as generally ignoring it. A Disallow written for the first does not constrain the second.
The same citation-crawler role at OpenAI. Both belong in the allow list if your goal is AI search visibility.
There is no Perplexity equivalent of GPTBot - Perplexity states neither of its bots trains foundation models, so there is no training token to block.
PerplexityBot FAQs
Does Perplexity use my content to train AI models?
Perplexity's crawler documentation states that neither PerplexityBot nor Perplexity-User trains foundation models. Their stated purpose is indexing and answering with citations, so there is no separate training token to opt out of.
Should I allow PerplexityBot?
For most publishers, yes. Perplexity attaches visible numbered citations to its answers, so being indexed is what makes you eligible for those referral clicks - and there is no documented training use to weigh against it.
Where do I get PerplexityBot's IP ranges?
Perplexity's current documentation points at a JSON feed on perplexity.com. If your verification code still fetches the older perplexity.ai path, check that it has not started silently failing.
Does blocking PerplexityBot also block Perplexity-User?
No. They are separate tokens with separate robots.txt groups, and Perplexity-User is documented as generally ignoring robots.txt anyway. Blocking the index crawler does not stop live user-triggered fetches.
Can PerplexityBot read your page right now?
Test any URL and see exactly what AI crawlers receive.