Perplexity-User
| Vendor | Perplexity AI |
| Type | Live retrieval / on-demand fetcher |
| robots.txt token | Perplexity-User |
| JavaScript rendering | Yes — uses a real browser context for some flows |
| Honors robots.txt | Yes (per vendor docs since 2024) |
| Vendor docs | docs.perplexity.ai/guides/bots |
User-Agent strings
Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; Perplexity-User/1.0; +https://perplexity.ai/perplexity-user)Purpose
The bot that fetches a URL because a Perplexity user just asked a question that requires it — including the "Pages", "Shop", and follow-up modes that retrieve fresh web data. Each request is tied to an active user query.
For ecommerce, Perplexity-User is the bot you actually want to allow: its requests are the ones that put your products in front of a customer at the moment they're shopping.
In our audit
Tested in v3 (rendered) at 15 sites. Received full content at walmart.com, samsung.com, and amazon.com (the 202 stub everyone gets). At amazon.co.uk it received the small stub — the trusted- crawler allowlist there excludes Perplexity-User.
How to allow / block
To allow Perplexity live retrieval:
User-agent: Perplexity-User
Allow: /To block:
User-agent: Perplexity-User
Disallow: /Quirks
- Distinct from PerplexityBot. A site can block PerplexityBot
(training/index) and still allow Perplexity-User (live, per-query).
- Uses a real headless browser for some flows; HTTP fetch for others.
- After the 2024 robots.txt controversy involving PerplexityBot,
Perplexity introduced this distinct token for user-driven fetches and committed to honoring robots.txt for both.