Perplexity-User
Fetches a page in real time when answering a user's question needs it. Documented as generally ignoring robots.txt.
Reference reviewed by Sona
| Operator | Perplexity |
|---|---|
| Powers | Live fetches for Perplexity answers |
| Purpose | User-triggered fetch |
| User-agent token | Perplexity-User |
| Respects robots.txt | No (user-triggered) |
Perplexity-User is the on-demand half of Perplexity's crawling. It fires when answering a specific user's question requires retrieving a specific page, rather than relying on what PerplexityBot has already indexed. Perplexity describes it as supporting user actions within Perplexity.
Perplexity states plainly that Perplexity-User generally ignores robots.txt, on the reasoning that a person asked for the page. Treat a Disallow for this token as a stated preference with no enforcement behind it. If you need the block to hold, it has to live at the server, CDN, or WAF layer.
As with PerplexityBot, Perplexity documents that this fetcher does not train foundation models. So the trade-off is narrow and worth naming precisely: allowing it means Perplexity answers questions about your page from the current page, and blocking it means Perplexity answers from whatever it already has, or from someone else's description of you.
Perplexity publishes a dedicated IP feed for this token, separate from PerplexityBot's. That separation is what makes edge enforcement practical - you can block unverified traffic wearing the name while allowing genuine fetches, or block the genuine ones precisely, without one rule spilling into the other.
How Perplexity-User behaves
- Single-page fetches triggered by one user's question, with no crawl pattern behind them.
- Documented as generally ignoring robots.txt, so Disallow rules for this token are advisory only.
- Has its own IP feed, separate from PerplexityBot's - do not verify one against the other's list.
- Perplexity documents that these fetches do not feed foundation-model training.
Full user-agent string
Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; Perplexity-User/1.0; +https://perplexity.ai/perplexity-user)
Allow Perplexity-User
A user asking Perplexity about your page gets an answer grounded in the live page, with a citation - and Perplexity states the content is not used for training.
User-agent: Perplexity-User Allow: /
Block Perplexity-User
You want to prevent all programmatic access from Perplexity, including user-initiated fetches - but enforce it in your WAF, since robots.txt is documented as generally ignored.
User-agent: Perplexity-User Disallow: /
Heads up: Perplexity-User does not reliably honor robots.txt. To enforce a block, combine the rule above with server- or firewall-level filtering of the user-agent.
How to verify Perplexity-User
Published IP range feed
Perplexity publishes a separate feed for this token, distinct from PerplexityBot's - verify against the perplexity-user file specifically. This is the practical foundation for enforcement: because robots.txt is documented as generally ignored, the only way to block real Perplexity fetches deliberately (rather than blocking anything that claims the name) is to match the published ranges at the edge.
Check an IP against this botCommonly confused with Perplexity-User
PerplexityBot indexes and honors robots.txt; Perplexity-User fetches live and generally does not. Same vendor, opposite enforcement story.
The closest analogue at OpenAI - also user-triggered, also documented as not necessarily bound by robots.txt.
The contrast case: Anthropic documents Claude-User as respecting robots.txt do-not-crawl signals, where Perplexity-User does not.
Perplexity-User FAQs
Can robots.txt block Perplexity-User?
Not dependably. Perplexity documents that this fetcher generally ignores robots.txt because the request is user-initiated. A Disallow records your preference; blocking requires server, CDN, or WAF filtering.
How do I block Perplexity-User properly?
Match the user-agent at the edge and scope the rule to the IP ranges Perplexity publishes for this specific token. That way you block genuine fetches rather than anything that borrows the name.
Is my content used for training if I allow this?
Perplexity states that neither of its bots collects content for foundation-model training. Allowing Perplexity-User grounds live answers in your page; it does not, per Perplexity's documentation, contribute to a training corpus.
Why do I see Perplexity-User but no PerplexityBot in my logs?
They do different jobs. Perplexity-User fires only when a user's question requires a live fetch of that URL, so it can appear on pages the index crawler has not yet reached, or on sites where PerplexityBot is disallowed.
Can Perplexity-User read your page right now?
Test any URL and see exactly what AI crawlers receive.