How to Show Up in Perplexity Answers (2026)

Perplexity answers with numbered sources. Showing up means your URL is one of them, or at least your brand is named with a path a human can follow. The vendor already published the crawl rules. Perplexity crawlers is the 2026 page to follow, not a Twitter thread about "GEO chunks."

Named without a source link is a mention. A numbered citation is the check this title asks for. Do not mash them, because a mention that never became a citation is a different problem from a citation that disappeared, and the fix for one does not fix the other.

Two agents, two jobs

PerplexityBot surfaces and links sites in Perplexity search results. Perplexity says it is not used to crawl for foundation-model training. Allow it in robots.txt and allow the IPs at perplexitybot.json. Full user-agent includes PerplexityBot/1.0 and a docs link. Changes can take up to 24 hours.

Perplexity-User fetches a page because a user asked something that needs that page, and may add a link. It is not the training crawler and not the automatic indexer. Because a user requested it, Perplexity says this fetcher generally ignores robots.txt. IPs: perplexity-user.json. If you needed to hide a draft, Disallow is the wrong tool for User. Use auth.

If your WAF "blocks AI," you can allow the bot in robots.txt and still 403 the JSON ranges. Their docs tell you to match user-agent and IP. They sketch Cloudflare custom rules (Security, WAF, custom rules, User-Agent contains PerplexityBot or Perplexity-User, and IP is in the published ranges, action Allow) and AWS WAF IP sets plus string matches. Update those ranges. They say the JSON is the source of truth and changes. They also say WAF changes take time to propagate, so watch logs.

Do not invent CIDR lists in a spreadsheet. Fetch the JSON, because a hardcoded range that went stale is how a real Perplexity fetch gets 403'd by a rule that was written to allow it.

The two fetchers also explain why a robots.txt change is not a complete lever for Perplexity. The bot respects it. The user fetcher generally does not. A publisher who disallows PerplexityBot and assumes that also covers the user fetcher is hiding half the surface and leaving the other half exposed, and the half that is exposed is the one driven by a reader asking a live question.

Pages worth retrieving

Same boring web hygiene: public HTML, internal links, the claim in text. Google's optimization guide is still the best public essay on unique versus commodity pages. Perplexity did not copy it. The instinct travels: do not be the seventh summary of the same listicle, because a model that has six summaries already does not need a seventh, and the seventh is the one that gets dropped.

There is no Perplexity Search Console. There is no special schema SKU that forces a numbered source. llms.txt is not a documented admission ticket on that crawler page.

Perplexity 22M+ monthly active users on the figures we cite from Promptwatch's sheet. Smaller than ChatGPT. Still a real answer engine with citations you can lose.

The size gap does not make Perplexity optional. A reader who asks Perplexity a buyer question gets a numbered answer and tends to read the sources. Losing that citation loses the reader at the moment they are closest to a decision. The audience is smaller than ChatGPT's, but it is an audience that clicks the footnote, and that is the audience a citation program is for.

Measure the answer, not your robots.txt intentions

Promptwatch paid plans include daily Perplexity UI checks. Citation analytics show whether they used you or a publisher, including Reddit and YouTube when those win. Visitor analytics catch referred sessions. Essential is $95/mo. Explore is ChatGPT-only, so it will not answer this title. Agent Analytics can log PerplexityBot, but crawler logs start on Professional at $245/mo with a 25M allowance.

Otterly.AI has Perplexity in the base four from $29/mo. Mention check. Possible week lag. No published crawler-log product on our listing. Profound Starter is ChatGPT-only. Wrong engine. Peec AI last.

The order of the named tools follows the same rule we keep across this site. Otterly and Peec stay last in any ranking that names them, because they are mention tools that stop at the chart. A mention tool has a real use as a cheap proof. It does not have a use as the system of record for a citation program, and putting it above a platform that stores the cited URL would mislead a reader who came here to pick one.

2026 order

Allow PerplexityBot. Whitelist current JSON IPs at the WAF.
Do not assume Disallow stops Perplexity-User.
One page that answers the query with a fact you own.
Track that query daily on a paid plan for two weeks. Note cited URL, not a vibe.
If you never appear, read bot logs before you hire a writer. A 403 is not a GEO content problem.

That is how you show up. Let the search bot in. Be citable. Keep a ledger. Perplexity already told you step one. Product: promptwatch.com.