Skip to content
Chalk Labs / AI SEO
How-to guide

How to Rank in Perplexity and Get Cited

By Rahil Jain, founder of Chalk LabsUpdated How we research

The short answer

Perplexity searches the web live for each question and cites the pages it finds. To get cited, allow PerplexityBot in robots.txt and in your firewall, write sections that answer one question each and stand alone, back every number with a source, update pages when facts change, and check Perplexity's citation lists weekly. Perplexity publishes no ranking-factor list.

  • Perplexity says to allow PerplexityBot in robots.txt and to permit its published IP ranges.
  • Perplexity splits pages into self-contained spans, so each section must make sense alone.
  • Most-cited-site tables disagree between studies. Plan around your own prompts, not a leaderboard.
  • Citations fade fast: one study found the median page lost half its share within 11 days of peaking.

How does Perplexity pick the pages it cites?

Perplexity searches the web in real time for each question, gathers material from sources such as articles, websites and journals, and attaches numbered citations to the answer 1. You are not competing for a place in a fixed list. You are competing at the moment someone asks.

Perplexity says its own search index tracks over 200 billion unique URLs 2. It breaks documents into self-contained spans, and each span can be retrieved and ranked on its own 2. It searches by keywords and by meaning, merges the two result sets, and uses stronger scoring models for the final order 2.

Perplexity does not publish a list of ranking factors, and we found none in the pages we read. Everything below comes from its crawler and index documents plus outside studies. Treat the outside studies as hints, not rules.

Do it step by step

0 of 8 done0%

  1. Add the block from the templates, keeping your private-path rules inside it. Perplexity recommends allowing PerplexityBot 3. Open yourdomain.com/robots.txt afterwards to confirm it loads and shows the new lines.

    Tip: A bot with its own group ignores the User-agent: * group, so repeat any Disallow lines you still need.

  2. Run the IP script from the templates. Paste its output into a custom rule in your CDN or firewall that matches the user agent and the address, with the action set to allow, as Perplexity's guide advises 3. Do the same in any second firewall layer your host adds.

    Tip: Name the rule clearly and add the date, so the next person knows why it exists.

  3. After 24 hours 3, run the log commands from the templates. You want PerplexityBot hits with status 200 on your key pages. A 403 or 429 means something is still refusing it. Compare the visiting addresses with the IP list.

    Tip: No hits yet is normal on a new or small site. Perplexity gives no schedule for crawling.

  4. Write ten questions your buyers ask, run each in Perplexity, and open the citation list. In the prompt log from the templates, record each cited domain and its page type: article, video, forum, directory or vendor page. The pattern tells you what to build.

    Tip: Use the free Query Fan-Out Explorer to see the sub-searches behind each prompt.

  5. Use the section template. Give each section a question heading, a first-sentence answer, one sourced number and one named quote or example. Check that every section still reads well if it is lifted out alone 2.

    Tip: Delete phrases like 'as mentioned above'. A lifted span cannot see 'above'.

  6. If Perplexity cites videos for your topic, plan a short video. If it cites directories, fix your listings there. If it cites vendor pages, write a clearer one. Do not copy a rival, but do meet the format buyers and the engine seem to prefer.

  7. Show the date you last checked the page's facts, and keep a private log of what changed. Update when prices, rules or numbers change. Do not change dates without changing the page 9.

    Tip: A quarterly fact check for pricing and policy pages is a sensible default. That is our suggestion, not a measured rule.

  8. Re-run the same ten prompts on the same day each week and log whether you are cited. Expect some pages to appear and vanish, because the median page lost half its citation share in 11 days in one large study 10. Judge the trend over a month.

Your ticks are saved in this browser only.

Copy-ready templates

robots.txt: allow PerplexityBot and Perplexity-User

text
# Perplexity's search bot and its user-triggered fetcher.
# A bot with its own group ignores the User-agent: * group,
# so repeat private-path rules here.
User-agent: PerplexityBot
User-agent: Perplexity-User
Disallow: /admin/
Allow: /

# Everyone else
User-agent: *
Disallow: /admin/

Replace /admin/ with your own private paths. Perplexity says changes can take up to 24 hours to apply, and that Perplexity-User generally ignores robots.txt anyway . Keep truly private pages behind a login.

Python 3: PerplexityBot IP ranges and a Cloudflare rule

text
#!/usr/bin/env python3
"""Usage: python3 perplexity_ips.py

Downloads Perplexity's published PerplexityBot IP ranges and prints them as
a plain list plus a ready-to-paste Cloudflare custom-rule expression that
matches BOTH the user agent and the address, as Perplexity advises."""
import json
import urllib.request

URL = "https://www.perplexity.com/perplexitybot.json"
req = urllib.request.Request(URL, headers={"User-Agent": "ip-list/1.0"})
with urllib.request.urlopen(req, timeout=20) as r:
    data = json.load(r)

cidrs = [p.get("ipv4Prefix") or p.get("ipv6Prefix") for p in data.get("prefixes", [])]
cidrs = [c for c in cidrs if c]

print("# Published:", data.get("creationTime"), "| ranges:", len(cidrs))
print("\n".join(cidrs))
print()
print("# Cloudflare custom rule. Set the action to Allow, as Perplexity's guide describes.")
print("# Re-run this script now and then: the published list can change.")
print('(http.user_agent contains "PerplexityBot" and ip.src in {' + " ".join(cidrs) + "})")

Save as perplexity_ips.py and run it. It reads Perplexity's published JSON file and prints the ranges and a rule expression. Other firewalls use a different syntax, so adapt the list. Perplexity's own page gives Cloudflare and AWS WAF steps .

Bash: what has PerplexityBot fetched from your site?

bash
LOG=access.log   # your log path. Use zgrep for rotated .gz files.

# 1. How many requests name each Perplexity bot?
for b in PerplexityBot Perplexity-User; do
  printf "%-16s %s\n" "$b" "$(grep -c "$b" "$LOG")"
done

# 2. Status codes PerplexityBot received (field 9 in the combined log format)
grep 'PerplexityBot' "$LOG" | awk '{print $9}' | sort | uniq -c | sort -rn

# 3. Which pages did it fetch, and what did it get?
grep 'PerplexityBot' "$LOG" | awk '{print $7, $9}' | sort | uniq -c | sort -rn | head -20

# 4. Which addresses visited? Compare with the list from perplexity_ips.py.
grep 'PerplexityBot' "$LOG" | awk '{print $1}' | sort | uniq -c | sort -rn | head -10

Works with the standard nginx and Apache combined log format: field 1 is the address, field 7 the path and field 9 the status. Other formats need different field numbers. Behind a CDN, make sure the log records the real visitor address.

Self-contained section template

markdown
## [Question a buyer would ask]?

[One or two sentences that answer it directly, naming the subject in full.]

- Evidence: [number or fact] ([source, year]).
- Example: [one concrete case, with the date or place].
- Expert: "[short quote]" ([name], [role]).

*Last checked on [date].*

Use one block per question. Name the subject in full each time, so the section still makes sense when lifted alone.

Prompt log (copy into a spreadsheet)

markdown
| Date | Prompt | Cited domains (in order) | Page type of each | We are cited? | Position | Notes |
|---|---|---|---|---|---|---|
| 2026-11-03 | [prompt 1] | | | yes / no | | |
| 2026-11-03 | [prompt 2] | | | yes / no | | |

Log the same ten prompts each week. Screenshots are your evidence, because citations change between runs.

Does Perplexity need permission to cite you?

In practice, yes. Perplexity says PerplexityBot is built to surface and link websites in its search results, and it recommends allowing PerplexityBot in robots.txt and permitting its published IP ranges 3. It adds that PerplexityBot is not used to crawl content for training AI foundation models 3.

A second name, Perplexity-User, fetches a page when a person asks about it. Perplexity says that fetcher generally ignores robots.txt, because a user requested the fetch 3. Rule changes can take up to 24 hours to take effect 3.

Firewalls need care. Perplexity advises a rule that matches both the user agent and the published IP ranges, with the action set to allow 3. Cloudflare alleged in August 2025 that Perplexity also used undeclared crawlers with a generic Chrome user agent, and said it had de-listed Perplexity as a verified bot 4.

Perplexity replied that Cloudflare had confused its traffic with unrelated requests from a cloud-browser provider 5. Neither side's claim is settled here. Read your own CDN's bot settings and logs instead of assuming either account.

What kinds of sources does Perplexity cite most?

Studies disagree, so do not plan around one table. Profound analysed 680 million citations from August 2024 to June 2025, and Reddit made up 6.6% of Perplexity's citations, YouTube 2.0% and Gartner 1.0% 6. Ahrefs studied about 953,500 Perplexity prompts in June 2025 and put YouTube at 16.1% mention share and Wikipedia at 12.5% 7.

The two studies use different measures and prompt sets, so their numbers cannot be compared. Both show that a few large platforms take a big share. Our reading: for most businesses, the realistic target is the narrow questions those platforms answer poorly.

Perplexity leans on Google rankings more than its rivals do. In Ahrefs' test of 15,000 long-tail prompts, 28.6% of Perplexity's cited URLs sat in Google's top 10, against about 12% for ChatGPT, Gemini and Copilot 8. Ahrefs notes that Perplexity uses its own crawler-based index, not Google's or Bing's 8.

Two studies of Perplexity's most-cited sources (data from mid-2025), and why they differ
StudyPeriod and sizeWhat it measuresTop sources on Perplexity
Profound 6680 million citations, August 2024 to June 2025Share of all citations in its datasetReddit 6.6%, YouTube 2.0%, Gartner 1.0%, Yelp 0.8%
Ahrefs 7About 953,500 Perplexity prompts, June 2025Mention share: how often the model chooses a domainYouTube 16.1%, Wikipedia 12.5%

Does freshness matter for Perplexity?

A little, but the evidence is thin. In Ahrefs' study of 16.975 million cited URLs, Perplexity's in-text citations averaged 1,166 days since publication and 993 days since the last update, against 1,416 and 1,047 for ordinary Google results in Ahrefs' table 9. Cited pages are slightly newer than organic ones, not brand new.

Perplexity's engineering post says a model decides whether to index a URL from its importance and likely update frequency, and that high-traffic documents are continually reindexed 2. It gives no recrawl intervals 2. Our reading: pages that change often and matter to many readers are revisited more.

Citations also fade. Profound tracked 883,000 pages across seven answer engines for a year and found the median page lost half its citation share within 11 days of its peak 10. It reports that Perplexity and Claude return to pages repeatedly 10. Publish on a rhythm, and update when facts change. Ahrefs notes that Google's John Mueller has warned against changing dates without changing the page 9.

How should you structure a page for Perplexity?

Make every section able to stand alone. Perplexity says it splits documents into self-contained spans that are retrieved and ranked individually 2. So each section should name its subject, answer in its first sentence, and never lean on 'as above'.

Add evidence a model can lift: a number with its source, a dated fact, a named expert quote. The GEO research paper tested nine content changes, including one run on live Perplexity.ai. Quotation addition scored about 21% to 22% above baseline on the paper's word-count measure; keyword stuffing scored about 9% below on that measure, though it did not hurt on the paper's second measure 11.

Write headings as the questions buyers ask, and keep answers in plain visible text. How to write content AI engines quote gives the full pattern, and the free Answer Block Checker tests an opening paragraph.

What did an AI assistant search for when we asked this question?

On 20 October 2026 a Claude agent with web search ran five searches for 'how to get my website cited in Perplexity', covering tactics, ranking factors, access, source studies and content habits.

One run on one engine is a sample, not a rule, and Perplexity's own system may split the question differently.

How do you check whether Perplexity cites you?

Ask the question and read the numbered citations. Run your ten prompts in Perplexity, open each citation list, and note whether your domain appears and in what position 1. Repeat weekly, because citations fade quickly 10.

Expect modest visits. In Ahrefs' September 2026 benchmark across 115,254 sites, Perplexity sent about 64,000 visitors, 0.01% of the total 12. StatCounter's chart for the same month puts Perplexity at 5.85% of worldwide AI chatbot market share, though the page does not say how it measures that 15.

Cloudflare's 2025 data show Perplexity with the lowest crawl-to-refer ratios, the number of bot requests per visitor referred, of the major AI platforms (under 200:1 from September 2025, though it spiked above 700:1 in March). Anthropic's ranged from about 25,000:1 to 100,000:1 later in the year, peaking at 500,000:1 earlier 13. On publisher sites over one month, Microsoft Clarity saw Perplexity visitors subscribe at 7 times the rate of direct and search visitors, a small, one-sector sample 14. Search your referrers for perplexity.ai to see your own numbers.

Which mistakes cost the most time?

Six mistakes account for most of the wasted effort. Each has a cheap fix.

Common Perplexity mistakes and fixes
MistakeWhy it hurtsFix
Blocking PerplexityBot to stop 'scraping'It is the bot Perplexity says surfaces sites in results 3.Block training bots separately, not this one.
Allowing the bot in robots.txt but blocking it in the firewallPerplexity advises a firewall rule on user agent plus IP 3.Add the rule and check your logs.
Allowlisting by user agent aloneAnyone can type a bot name, so the name proves nothing.Match the published IP ranges too 3.
Stuffing keywordsIt scored about 10% below baseline in the paper's live Perplexity test 11.Add sourced numbers and named quotes.
Planning around a 'top cited sites' chartTwo studies disagree on the leaders 67.Map the sites cited for your own prompts.
Changing dates without changing the pageAhrefs notes Mueller warned against it 9.Update dates only with real edits.

When is Perplexity not worth the effort?

Wait if your buyers do not use it. Check your own analytics for perplexity.ai referrals over the last 90 days, and ask recent customers which AI tools they used. If you see none, put the first effort into Google and ChatGPT, where the traffic is larger 12.

Also hold back if your pages carry content you cannot publish openly, such as paid research. Perplexity-User may fetch a page a person asks about even where robots.txt says no 3, so use logins for private material. For a side-by-side view, read how ChatGPT, Perplexity and Gemini cite sources.

What an AI assistant searches when it answers this

We gave the question below to AI search tools and recorded the web searches they ran behind the scenes. This is called query fan-out. Each phrase is a page you could publish to be found and cited.

The prompt

“how to get my website cited in Perplexity”

The searches it ran
  • how to get cited in Perplexity
  • Perplexity ranking factors sources citations
  • PerplexityBot robots.txt allow
  • Perplexity most cited domains study
  • Perplexity optimization content freshness structure tips

Observed October 20, 2026 using Claude web search. Claude agents were told to run 6 to 10 searches, so that count is our setting, not Claude's own choice. Assistants differ and change often, so treat this as a sample, not a rule. See how we collect it.

Frequently asked questions

Should I block PerplexityBot?

Only if you do not want to appear in Perplexity's search results. Perplexity says PerplexityBot exists to surface and link websites in those results and is not used to crawl content for AI foundation models 3. If your worry is training use, that is a different bot decision.

Does Perplexity obey robots.txt?

Perplexity says PerplexityBot follows your rules, while Perplexity-User generally ignores robots.txt because a person asked for the fetch 3. Cloudflare alleged in 2025 that Perplexity also crawled in ways that evaded no-crawl rules, and Perplexity denied it 45. Use a login for pages that must stay private.

Is there a Perplexity ranking factor list I can follow?

No. Perplexity does not publish one, and we found none in its help or engineering pages. It describes real-time search, span-level retrieval, hybrid keyword and meaning search and a final reranking step 12. Outside studies offer hints on source types, but they disagree 67.

How often should I update pages for Perplexity?

Update when facts change, and review key pages on a schedule you can keep. Perplexity says high-traffic documents are continually reindexed 2, and one study found citations fade within weeks 10. Do not change dates without changing the page, because Ahrefs notes Google has warned against it 9.

Do I need to be on Reddit or YouTube to be cited?

Not always, but both appear high in studies of Perplexity's sources 67. Check which page types Perplexity cites for your own prompts before you invest. Read how to get mentioned on Reddit without getting banned before you post, because Reddit communities punish promotion.

About the author

Rahil Jain

Rahil Jain is an AI SEO (answer engine and generative engine optimisation) consultant and co-founder of Chalk Labs.

LinkedIn · Method and credentials

Sources

  1. How does Perplexity work? · Perplexity Help CenterOpened October 20, 2026. Supports: Perplexity interprets the question, searches the internet in real time, gathers information from sources such as articles, websites and journals, compiles an answer and attaches numbered citations.
  2. Architecting and evaluating an AI-first Search API · Perplexity ResearchOpened October 20, 2026. Supports: Published 25 September 2025: index tracks over 200 billion unique URLs; documents decomposed into self-contained spans that are individually retrieved and ranked; lexical and semantic retrieval merged; cross-encoder reranking; a model decides indexing from importance and likely update frequency; high-traffic documents continually reindexed; no recrawl intervals given.
  3. Perplexity Crawlers · PerplexityOpened October 20, 2026. Supports: PerplexityBot surfaces and links sites in results and is not used to crawl for AI foundation models; Perplexity-User generally ignores robots.txt; allow PerplexityBot and its published IP ranges; match user agent and IP in firewall rules; up to 24 hours for changes.
  4. Perplexity is using stealth, undeclared crawlers to evade website no-crawl directives · CloudflareOpened October 20, 2026. Supports: Published 4 August 2025: Cloudflare's allegation that Perplexity used undeclared crawlers with a generic Chrome-on-macOS user agent, and that it de-listed Perplexity as a verified bot. An allegation, not a finding.
  5. Agents or bots? Making sense of AI on the open web · PerplexityOpened October 20, 2026. Supports: Published 4 August 2025: Perplexity's response, saying Cloudflare confused Perplexity with 3-6 million daily requests of unrelated traffic from BrowserBase.
  6. AI Platform Citation Patterns · ProfoundOpened October 20, 2026. Supports: Published 5 June 2025: 680 million citations from August 2024 to June 2025; on Perplexity, Reddit 6.6%, YouTube 2.0%, Gartner 1.0% and Yelp 0.8% of total citations in the dataset.
  7. The 10 Most Mentioned Domains for ChatGPT, Perplexity, and AI Overviews Across 78.6M Searches · AhrefsOpened October 20, 2026. Supports: Published 11 June 2025: about 953,500 Perplexity prompts; YouTube 16.1% and Wikipedia 12.5% mention share; mention share measures how often the model chooses a domain.
  8. Only 12% of AI Cited URLs Rank in Google's Top 10 for the Original Prompt · AhrefsOpened October 20, 2026. Supports: Published 11 August 2025: 15,000 long-tail prompts; 28.6% of Perplexity's cited URLs in Google's top 10; about 12% for ChatGPT, Gemini and Copilot; Perplexity uses its own crawler-based index.
  9. Do AI Assistants Prefer to Cite Fresh Content? · AhrefsOpened October 20, 2026. Supports: Published 28 July 2025: 16.975 million cited URLs; Perplexity in-text citations 1,166 days since publication and 993 since update; organic 1,416 and 1,047; the author notes Mueller's warning against changing dates without changing content.
  10. The half-life of an AI citation is 11 days · ProfoundOpened October 20, 2026. Supports: Published 30 September 2026: 883,000 pages across seven answer engines from September 2025 to September 2026; median page loses half its citation share within 11 days of peaking; Perplexity and Claude return to pages repeatedly.
  11. GEO: Generative Engine Optimization · arXiv (Aggarwal et al.; KDD 2024)Opened October 20, 2026. Supports: Nine content methods tested; in a live Perplexity.ai test, Quotation Addition scored about 22% above baseline and keyword stuffing about 9% below on the position-adjusted word count measure; main results come from the authors' own test engine.
  12. AI vs Search traffic benchmark · AhrefsOpened October 20, 2026. Supports: September 2026 across 115,254 sites: Perplexity about 64,000 visitors (0.01% of traffic); ChatGPT 4.3 million (0.34%).
  13. Cloudflare Radar 2025 Year in Review · CloudflareOpened October 20, 2026. Supports: Crawl-to-refer ratio defined as bot requests compared with referrals; Perplexity had the lowest ratios of the major AI platforms, below 200:1 from September 2025; Anthropic as much as 500,000:1. Data 1 January to 2 December 2025.
  14. AI Traffic Converts at 3x the Rate of Other Channels (Study) · Microsoft ClarityOpened October 20, 2026. Supports: Over 1,200 publisher and news sites, one month of conversion data: Perplexity visitors had a subscription rate 7 times direct and search; AI referrals under 1% of traffic.
  15. AI Chatbot Market Share Worldwide · StatCounter Global StatsOpened October 20, 2026. Supports: September 2026: Perplexity 5.85% worldwide AI chatbot market share; the page does not state its measurement method.
How this page was made. Built on Rahil Jain's AI SEO method and drafted with AI research assistance. Every statistic links to the page we read it on (opened October 20, 2026 or nearby). Published October 20, 2026; last updated October 20, 2026. Chalk Labs sells AI SEO services, so we have a commercial interest in this topic. Spotted an error? Tell us and we will fix it.