How to Get Cited by Perplexity
Perplexity shows its sources, which makes it the easiest engine to debug — you can see exactly which pages it preferred over yours.
Updated October 24, 2026 · 7 min read · part of the GEO guides
Every Perplexity answer carries numbered citations. That transparency is a gift: instead of guessing why you were not picked, you can open the answer, read the cited passages, and compare them to your own page. Most of the work is closing that gap.
First, check that both Perplexity crawlers can fetch you
Perplexity operates two user agents and they do different jobs:
- PerplexityBot — the indexing crawler. If this is blocked, none of your pages exist for Perplexity.
- Perplexity-User — fetches a page live when a user asks about it. Blocking this means you can be cited only from an existing index copy, or not at all.
Both are covered in the crawler directory, with the exact robots.txt group syntax in the robots.txt guide. Verify your current verdict per bot with the Auditor — a blanket Disallow: / for AI bots is the single most common reason a site is invisible here.
Also rule out server-side blocks: a WAF or bot-protection rule that never appears in robots.txt will return 403 to PerplexityBot while your browser sees a perfectly healthy page. Fetch your page as the bot with the live scanner to catch that.
Write the sentence you want quoted
Perplexity cites passages, and a cited passage is usually the one that answers the question directly, early, and without needing the surrounding context. Practical rules:
- Put the direct answer in the first sentence under the heading.
- Include the concrete number, date or definition that the query is asking for.
- Keep the passage self-contained — avoid pronouns that refer to a paragraph three sections earlier.
- Name the thing you are talking about, not just the category.
Compare your page against a cited competitor at sentence level. If their paragraph answers the question in 25 words and yours takes 120, they win the citation.
Give it a comparison table
Perplexity frequently answers "X vs Y" and "best Z" queries with a table. Pages that already contain a real HTML <table> with consistent columns are far easier to summarize than pages where the same information is scattered through prose. If you publish comparisons, publish them as tables with explicit column headers and units.
Make facts verifiable from the page
Citations are chosen partly on trust. A statement with a linked primary source, a date, and an attributable author is safer to cite than an unlinked assertion. Include:
- A visible updated date, mirrored in
dateModified. - A named author or organization with a real page behind it.
- Outbound links to the primary source of any statistic you quote.
The schema guide covers the markup side; the visible side matters just as much.
Keep the crawl path shallow
If your best page is four clicks from the homepage and linked from nowhere, it is crawled rarely and re-indexed slowly. Build topic hubs, link related pages bidirectionally, and expose your strongest pages in llms.txt — the llms.txt guide explains what that file does and does not do. The generator tab on the homepage builds one from your sitemap in seconds.
Run the loop weekly
The honest workflow is: ask Perplexity your ten most commercially important questions, record which URLs it cites, and diff those against your own pages. Anything a competitor does structurally that you do not — answer-first paragraphs, tables, dates, sources — is a concrete task. Repeat weekly; re-indexing takes days to weeks, so single-day checks tell you nothing.
Check the whole site, not one URL
The free scanner tests one page against the crawlers and the on-page signals. If you want all of it at once — up to 50 pages, per-page bot access, noindex traps, broken links and a prioritized P0–P3 list — the Pro Deep Report is $49 one-time and returns a private report URL you can hand to whoever maintains the site.
Frequently asked questions
What is the difference between PerplexityBot and Perplexity-User?
PerplexityBot crawls and indexes pages for Perplexity's index. Perplexity-User fetches a specific page live when a person asks a question during a session. Blocking PerplexityBot removes you from the index; blocking Perplexity-User prevents on-demand fetches, which can still cost you citations.
Does Perplexity respect robots.txt?
Yes, both Perplexity crawlers respect robots.txt, and Perplexity publishes documentation for both user agents. If you are seeing 403s in your logs while robots.txt allows them, the block is happening at your CDN or WAF, not in the text file.
How do I see which sources Perplexity cited instead of me?
Open the answer and read the numbered citations, then open those pages and compare their structure to yours at the paragraph level. Look specifically for answer-first opening sentences, concrete numbers, tables and linked sources — those are the usual differences.
Can I pay to appear in Perplexity answers?
No. Citations are selected by the retrieval system from indexed and live-fetched content. There is no submission form and no paid placement; the controllable levers are crawl access, passage structure, entity clarity and freshness.
Keep reading
- How to Get Cited by Microsoft Copilot
Copilot answers come from the Bing index, so Bingbot access and IndexNow freshness matter more than most GEO advice admits. A practical checklist for Bing, Copilot and Edge surfaces.
- The 12-Point AI Visibility Audit
A practical 12-point audit you can run in under an hour: crawler access per bot, bot-vs-browser fetches, passage structure, structured data, llms.txt, noindex traps and measurement.
- What is Generative Engine Optimization (GEO)?
GEO is the practice of making your site readable, extractable and citable by AI answer engines. What it covers, why it matters, and the first five steps to take.
Crawler profiles: GPTBot · OAI-SearchBot · ClaudeBot · PerplexityBot · Google-Extended · CCBot — or see the full AI crawler directory.
See what AI engines can see on your site
Audit your robots.txt against 61 AI crawlers in seconds — free, no signup, runs entirely in your browser.