Track your visibility in Perplexity

本文目前只有英文版。站点其余部分已汉化，这篇内容尚未翻译。

Perplexity is the engine where measuring AI visibility is least ambiguous, and the reason is architectural rather than lucky: it was built to show its work. Every answer arrives with the pages it drew from attached, which means the question "was this domain cited" has a factual answer that does not depend on anyone's reading of the prose.

That makes it a good place to explain what a citation check is actually doing — because here, none of the machinery is compensating for a missing signal.

How the probe works

Each tracked query goes to the Perplexity sonar API, which runs a live web search and returns both an answer and a structured list of the sources behind it. We keep up to twelve of those sources per query, verbatim, so every verdict stays auditable long after the run.

There is no forcing step needed. With the other engines a real design decision has to be made about how to guarantee the model actually retrieved something rather than answering from memory. Perplexity's normal operating mode is retrieval, so the probe is close to the plain question a user would ask.

What "cited" means here

A query counts as cited when one of the returned source URLs resolves to your domain. We reduce each URL to its host, lowercase it, drop a leading www., and count it only when the host is your domain exactly or a subdomain of it.

The same three consequences apply here as everywhere else in TriRank, and they are worth restating because they are what separates a real number from a flattering one:

• Prose mentions do not count. If the answer names your brand with no source behind it, we record not cited.
• Being written about is not being cited. A citation to somebody else's review of your product is that site's citation, not yours. It still appears in your source list, which is where it is useful.
• Subdomains count as you. Documentation and help-centre subdomains are part of your site.

What this cannot tell you

A citation is not a click. Perplexity displays sources more prominently than most engines, which genuinely helps — but the answer is designed to satisfy the user without a click, and often does. This is the zero-click dynamic applied to a new surface. What a citation tells you is that you were part of the answer a real question received. What it does not tell you is how many people went on to visit.

One reading is noise. Live retrieval means the same question asked twice can surface different pages. A single check is not a measurement; the weekly series is. Any tool that hands you a one-off screenshot is selling you a coin flip with a confident caption.

Ranking here does not follow ranking there. The pages Perplexity retrieves for a question frequently are not the pages ranking on the first page of Google for the same words. That is not a bug in either system — they are optimising for different things — but it does mean your SEO position is a poor predictor of your position here, in both directions.

Query wording changes everything. Answer engines fan a question out into several internal searches, so a small change in phrasing can pull a completely different source set. This is why the queries you track are a decision worth making carefully rather than a list to fill.

There is no position number here, and there is not one to have. A results page has ten slots, so a rank is a row you either occupy or you do not. A synthesized answer is prose that names a few brands and links a few pages, and there is rarely a numbered position to track inside it. What this check reports is therefore presence and citation, not a row number, and nothing on this page converts into one. Where you would want a rank, the measurement doing that job is the trend — how often this engine cited you across weeks of the same questions, rather than where you placed inside any single answer. The word "rank" is still settling in this market: we use it in the sense set out under AI search ranking, and AI search rank tracking covers what tracking it involves.

There is no published cross-brand benchmark for this engine. What runs on your account is your own sweep: your tracked queries, through this engine, on the same weekly cadence as the other four. What does not exist is a public study across many brands that would let you ask how this engine behaves in general — how often it cites anyone, or which kinds of sites it tends to reach for. No number on this page comes from one, and we are not putting a date on when there will be one. The distinction is worth stating plainly, because "we do not measure this engine" and "we have not published a study about this engine" are different claims and only the second is true.

What to do with the result

The source list is the actionable part. For any query where you are absent, it tells you precisely which pages won the retrieval — and that list is the brief.

Two things tend to distinguish the pages that get retrieved. They answer the question early and explicitly rather than working up to it, and they make claims a machine can carry away intact: specific, attributed, and dated. Vague, hedged prose survives a human skim and does not survive summarisation.

If you want to see the current source set for your own domain without setting anything up, the free audit runs this same probe and shows you what came back. And if you are weighing this engine against the others, Perplexity vs ChatGPT covers how the two differ in what they retrieve and how they present it.