How to Get Your AI Tool Cited by Perplexity in 2026
Of all the answer engines, Perplexity is the one where you can stop guessing. Every answer carries numbered citations linked to real URLs, so "are we in the answer?" is a count rather than an argument. This is what its crawler needs, what quietly blocks it, and what to fix when the count is zero.
For makersStep 5 is a listing on the kind of page these engines cite. That part is free.
List my tool — free →Quick Answer
In order: let the crawler in, put your facts in text, and get onto the third-party pages that answer category questions. In that order, because the last one cannot help a site the crawler has never successfully read.
The single most common cause of a zero we see is not content. It is a managed bot rule on a CDN, blocking a crawler the site owner never decided anything about.
A number we will publish about ourselves
Anyone selling AI-search visibility should have to show their own. We track eight unprompted category prompts — questions where a buyer never names us — and we are named in zero of them. The scores that look better than that, ours included, get flattered by prompts that mention the brand, which is not the test.
We publish that because it is the honest frame for everything below: this is a slow, countable, mostly-technical game, and the vendors quoting you a visibility percentage without publishing their own prompt set are quoting you the flattering half.
The six checks, in order
1.Allow PerplexityBot in robots.txt, explicitly
Perplexity's own crawler documentation asks for two things: allow the PerplexityBot user agent in robots.txt, and permit requests from its published IP ranges. A blanket Disallow on unknown agents covers it by accident, and so does a robots.txt that was written before answer engines existed.
How to check: Fetch https://yourdomain.com/robots.txt in a private window and read it as a stranger would. If PerplexityBot is not named, it is relying on your default rule.
2.Check whether your WAF is blocking it without you
This is the one that catches people. Cloudflare de-listed Perplexity as a verified bot and added managed heuristics that block its undeclared crawling. If you run a managed bot ruleset, the block can be on by default — you did not choose it, your robots.txt says nothing about it, and your logs are the only place it shows.
How to check: Filter your CDN or origin access logs for the PerplexityBot user-agent string over the last 30 days. Zero requests on a site that has been live for months is a block, not a lack of interest.
3.Understand what a block actually costs you
Perplexity's help centre states that a page disallowed in robots.txt is not indexed for its full or partial text, but the domain, headline and a brief factual summary may still be indexed. That is the worst of both outcomes for a tool: you can be described without the detail that makes anyone choose you, and you cannot correct it.
How to check: Ask Perplexity a question your product is the answer to. If you appear as a bare name with no specifics while competitors get pricing and feature detail, you are in the summary-only state.
4.Put the answerable facts in text, not in images or JavaScript
Pricing inside a screenshot, feature tables rendered client-side after a hydration pass, integration lists behind a tab that only exists after a click — a crawler gets none of it. Pricing and 'does it integrate with X' are the two most common follow-up questions about any tool, and they are the two most commonly unreadable.
How to check: curl your pricing page and grep the output for your own prices. If they are not in the HTML, they are not in the answer.
5.Get onto the third-party pages that get cited
Answer engines resolve 'which tool should I use for X' against pages that compare several tools, not against one vendor's homepage. That is why directory listings, category roundups and comparison pages do disproportionate work here: they are the shape of the question. A vendor page can confirm a choice; it rarely creates one.
How to check: Ask the engine your category question and read the cited URLs. Count how many are vendor sites versus third-party lists. Then check whether you are on the lists.
6.Measure it as a count, not a feeling
Perplexity is the easiest engine to measure on precisely because every answer ships numbered, linked citations. You can count your appearances per prompt directly, with no inference. Pick ten prompts a buyer would actually type, run them on a fixed schedule, and record whether you were cited. That is a number that can move.
How to check: Write the ten prompts down before you run any of them. Prompts invented after seeing the results will flatter you.
Three things that do not move the count
- Publishing more pages on your own site. Volume is not the lever here. An answer engine cites the page that answers the question, and forty thin pages do not answer it better than one that does.
- Keyword-stuffing an "AI" angle onto every page. These engines are summarising your page for a person who asked a specific question. Padding makes the summary worse, not longer.
- Treating a zero as proof that nobody is looking. A zero means either you were not readable or you were not on the cited pages. Both are fixable, and neither is a verdict on demand.
Do the crawler work yourself. The listing part is free.
Steps 1 through 4 are yours and nobody can do them for you. Step 5 — being on a third-party page that answers a category question — starts with a free listing here, which is exactly the shape of page these engines cite when someone asks which tool to use.
More guides for makers
Affiliate disclosure: Some links on this page are affiliate links. If you sign up through them, AISO Tools may earn a commission at no extra cost to you. This never affects our rankings or reviews.
📬 Get the best new AI tools delivered weekly
One concise email with fresh launches, trending picks, and featured standouts.
Join thousands of professionals who discover the best AI tools every week. No spam — unsubscribe anytime.