Reachable as Perplexity-User: what the check tests and how to fix it

Weight 0.5. Fetches your page once with the Perplexity-User string. A 2xx with a non-empty body passes; other statuses, timeouts and errors fail.

Agent Protocols · Weight 0.5 in the w2agent Score

How common is it?

93% of 1219 measured sites where this check applied pass.

Latest audit of each site under rubric v3.1. Sites where the check did not apply or could not be measured (for example, no llms.txt for the llms.txt checks) are left out. A partial result does not count as a pass.

What this check tests

The audit requests the page you submitted once with the Perplexity-User string, the second of its three scored reach requests. The rule that turns the response into a result is shared with ChatGPT-User and Google-Agent; read the first of those for the full outcome table. In short, a 2xx status with a non-empty body passes at 100, any other status, an empty body, a timeout or a network error fails, and a missing record or an exhausted time budget is unmeasured and not scored.

The header sent is:

User-Agent: Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; Perplexity-User/1.0; +https://perplexity.ai/perplexity-user)

That matches the string in Perplexity's crawler documentation. Only the response status and whether the body has any bytes are used. The check does not recognise challenge pages, does not follow robots.txt and compares nothing against the audit's own crawler.

Where it sits in the queue matters on slow sites. All three agent fetches run one after another within a shared budget, the smaller of three timeouts and 15 seconds, checked before each fetch. Perplexity-User runs second. With the standard profile (10-second timeout, 15-second budget) the first fetch cannot use up the budget, so this agent is not skipped for time; only the third one can be.

For how this differs from AI agent access (robots.txt only) and edge agent access (tracked, with a control comparison and challenge detection), see the ChatGPT-User page.

Why it matters to agents

Perplexity describes Perplexity-User as supporting user actions within Perplexity: when a person asks a question, it might visit a page to answer and link to it. It says the agent is not used for web crawling or to collect training content, and that because a user requested the fetch it generally ignores robots.txt rules. PerplexityBot is the separate token for search indexing.

Two consequences follow. A robots.txt block is not the reliable way to stop or allow this agent, so a fail here points at your server or firewall. And Perplexity itself warns that a web application firewall may need explicit allow rules for its bots, which is the case this check detects, since a block by user-agent shows as a 403 or 429.

As with the other reach checks, the audit sends the string from its own address, so a rule that also verifies Perplexity's published address list can treat it differently from the real agent.

How to fix it

Perplexity's documentation gives steps for two firewalls. Both combine the user-agent with its published address list:

  • Cloudflare: in the dashboard go to Security, then WAF, then Custom rules. Create a rule where User Agent contains PerplexityBot or Perplexity-User, and IP Source Address is in the ranges from Perplexity's official endpoint. Set the action to Allow.
  • AWS WAF: create IP sets from the official endpoints, add string-match conditions for the two user-agent strings, create allow rules combining them, and attach them to the Web ACL with higher priority than your blocking rules.

The address lists are JSON files, one for Perplexity-User and one for PerplexityBot. Perplexity advises fetching them on a schedule rather than copying them once, because they change.

Then re-run the audit. Rate limits cause the same symptom as blocks, so check for a 429. If you use a different CDN, apply the same pattern: allow by user-agent and address range together.

If you want this agent blocked, a failing result is the correct outcome and you can ignore it. See robots.txt for AI bots for what robots.txt can and cannot do here.

Examples from measured sites

Up to 10 of the highest-scoring sites in each group, from recent audits. Failing includes partial results.

Passing

Failing or partial

Sources

Verified against these sources on

Does your site pass this check?

Check your own site →