The Perplexity citation playbook: how to get cited in 2026

To get cited by Perplexity, make sure PerplexityBot can crawl you, because Perplexity answers from an index that its own crawler builds, then give it passages worth quoting: a direct answer near the top of the page, specific facts with sources, and a date that shows the information is current. Perplexity lists its sources on every answer, so success is easy to check: ask it the questions your buyers ask and see whether you are in the list. It is also worth the effort. By Cloudflare's 2025 measurements, Perplexity sends far more visits back per page it crawls than OpenAI or Anthropic.

How Perplexity finds sources

Perplexity documents two user agents. PerplexityBot is "designed to surface and link websites in search results on Perplexity", and Perplexity says it "is not used to crawl content for AI foundation models." It obeys robots.txt. Perplexity-User handles live fetches when a user's question needs a specific page, and Perplexity says it "generally ignores robots.txt rules" because a person initiated the request.

That split matters for strategy. PerplexityBot builds the index Perplexity searches when it answers, so blocking it removes you from consideration in the ordinary case. Since it is not a training crawler, blocking it buys no protection from model training either. For almost every business that wants to be found, the right setting is to allow it.

Why Perplexity deserves its own attention

Perplexity is a smaller audience than ChatGPT or Google, but it returns more of its crawling as traffic. Cloudflare measured crawl-to-referral ratios across its network from January to July 2025. In July, Perplexity crawled about 194 pages for every visit it referred, against roughly 1,091 for OpenAI and 38,065 for Anthropic. Perplexity's ratio worsened over those months, from 54 to 1 in January, but it remained the most generous of the three by a wide margin. Its answers put numbered source links next to the text, which is part of why its users click through.

Perplexity has also built a commercial relationship with publishers. In July 2024 it launched a publishers program that shares advertising revenue when a participating publisher's content is cited, with early partners including Time, Fortune, and Der Spiegel, and expanded it later that year. For publishers, being cited by Perplexity can be revenue as well as reach.

The six-step playbook

1. Allow PerplexityBot, everywhere it matters

Check robots.txt for a PerplexityBot group, and remember that a named group replaces your wildcard rules for that bot entirely. Then check your CDN or firewall: bot-protection defaults can return a challenge page or a 403 error to PerplexityBot even when robots.txt allows it. Our AI crawler directory walks through both.

User-agent: PerplexityBot
Allow: /
Disallow: /admin/

User-agent: *
Allow: /
Disallow: /admin/

2. Put the answer in the raw HTML

Like the other AI crawlers, PerplexityBot reads the HTML your server returns and does not run JavaScript. Content that appears only after client-side rendering is invisible to it. See which rendering modes AI crawlers can read.

3. Lead every section with the answer

Perplexity answers by quoting and summarizing short passages. A section that opens with its answer, names its subject, and includes a sourced specific is easy to cite; a section that warms up for three sentences is not. The GEO research found that adding statistics, quotations, and citations to content raised its visibility in generated answers by up to about 40 percent, and that keyword stuffing lowered it.

4. Show that the information is current

Many Perplexity questions are about things that change: prices, product comparisons, regulations, and recent events. Put a visible "last updated" date on pages where currency matters, keep it honest, and refresh figures when they change. A stale number is an easy reason for any engine to prefer a fresher source.

5. Be present in the sources Perplexity already cites

For most questions, a few domains appear again and again in Perplexity's source lists: industry publications, review sites, forums, and video. Ask Perplexity your buyers' questions, note those domains, and work on being accurately represented there. Our glossary calls this concentration the citation oligarchy.

6. Measure with repetition, not screenshots

Perplexity's answers vary from one run to the next, like every AI assistant's. SparkToro found that ChatGPT, Claude, and Google's AI almost never returned the same list of brands twice for the same prompt. Ask each important question several times, over several days, and track the share of answers that cite you. One appearance is a data point; a rate is a result.

Which pages to work on first

Perplexity users tend to ask research questions: comparisons, explanations, and "what is the best option for" questions. Start with the pages that answer those for your business, because they are the ones a buyer's question can land on.

• Comparison pages that set out honestly how your option differs from the alternatives, including where an alternative is the better fit.
• Pricing and cost pages that state real numbers or ranges, with what changes them.
• Explanatory guides on the core questions in your field, written by someone with first-hand experience.
• Original data, even modest: a survey of your customers, benchmark results, or anonymized figures from your own work. Data is the thing other sources cannot copy, and it is what engines quote.

Notice what is not on the list: thin blog posts written to target a keyword. If a page would not help a person who asked the question directly, it will not help Perplexity answer it either, and it competes with your better pages for the same citation.

A 30-minute Perplexity check

1. Write ten questions a buyer would ask before choosing a provider like you.
2. Ask each one in Perplexity and record the cited sources, noting whether you appear and in what position.
3. Fetch one of your key pages as PerplexityBot with curl and confirm you get a 200 response with the content in the HTML.
4. List the third-party domains that appear in more than two answers. Those are your outreach targets.
5. Repeat the questions next week. Treat anything that appears only once as a lead, not a finding.

A note on stealth crawling

In August 2025 Cloudflare reported that when Perplexity's declared crawler was blocked, pages were fetched by an undeclared crawler presenting as a generic Chrome browser, at 3 to 6 million requests a day. Perplexity disputed the report. If your goal is visibility, this is irrelevant: you are allowing the crawler anyway. If you need to keep Perplexity out, robots.txt alone may not be enough, and the enforcement point is your server or CDN.

How SEO for AI Agents measures this

The AI crawler access check reads your robots.txt for PerplexityBot and Perplexity-User and reports the exact directive that applies, then sends a real request with PerplexityBot's published User-Agent and records the HTTP status your server returns. That catches the CDN challenge or 403 that a robots.txt review would miss. The glossary entry on PerplexityBot covers its history.

For the outcome, we put your buyers' questions to Perplexity several times and record how often you are cited, with the verbatim answer as the receipt. The source of citation check groups the sources Perplexity cited instead by domain, which turns step five of the playbook into a ready-made target list. The engine divergence view shows whether Perplexity treats you differently from the other engines, which is common.