What is Perplexity-User?
A Perplexity-User request comes from a live action in Perplexity. When someone asks a question, Perplexity may open a web page to produce an accurate answer and include a link to that source. The agent handles that retrieval at the time of the question.
This is not Perplexity's general indexing crawler. Perplexity-User does not roam the web automatically or build the search corpus in advance. A hit in your logs means a user request caused Perplexity to fetch the page, not that PerplexityBot completed a routine crawl.
That distinction changes what the traffic means for AI visibility. Letting Perplexity-User reach a page can make the page usable in the answer being generated at that moment. It does not by itself place the site in Perplexity's broader search results. PerplexityBot is the agent documented for surfacing and linking sites there.
Perplexity says Perplexity-User is not used to collect content for training AI foundation models. Its role is retrieval for a user, so there is no training opt-out to manage through this token. Search crawling and user retrieval remain separate controls.
Because the fetch follows a direct user request, Perplexity says this agent generally ignores robots.txt. The metadata therefore treats support as partial. A Perplexity-User rule may record your preference, but it should not be your only barrier when access must be denied.
The documented header is Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; Perplexity-User/1.0; +https://perplexity.ai/perplexity-user). Perplexity publishes current addresses at https://www.perplexity.com/perplexity-user.json and recommends checking the IP together with the header. Its crawler guide is the operator source for both identifiers.
