Perplexity visibility
Perplexity visibility: TrustInsider asks Perplexity the questions in your panel through its official API, always with web search, and for each answer keeps mentions, recommendations, search results, country, city and model.
A single mode
Perplexity is an answer engine: every question goes through a web search, and the measurement respects that. There is no “from memory” reading to compare with the others: Perplexity’s cell is compared only with the search cells of the other assistants.
For Perplexity, a request without search is not even prepared: the system refuses it before sending it, instead of presenting an answer born from a search as a reading of the model.
The steps of a measurement question sent to Perplexity: the market that goes with the search, the results returned, the answer and the outcome that decides whether it enters the count.
The question leaves with the market’s country and, if the panel is by city, with the city.
Perplexity returns the list of search results used to answer.
The text is read for mentions, recommendations and competitor names.
Valid, refused or incomplete: only valid answers enter the count.
Search results
Together with the answer, Perplexity returns the search results it used. TrustInsider keeps them with their address and domain, so you see whether your site is among the pages read and which third-party sites describe your product.
The results are the ones Perplexity declares in the answer, never rebuilt from the text. A site that appears in the text without being among the results stays a mention: the two columns are not mixed up.
A sample Perplexity answer about fictional brands: the recommendation of the product, the brand’s support page among the results, the competitor named and an enthusiasts’ forum among the pages read.
QuestionCan sports earbuds last a marathon in the rain?
Perplexity · sample answer about fictional brands
The Ilmarde Audio Quessio 2 resist splashes and last beyond a long race, according to the support page on ilmarde-audio.example. Varolde Audio’s earbuds have more battery life, but runners on an enthusiasts’ forum report some problems with humidity.
The answer presents the brand’s product as suited to the question: it counts as a recommendation.
The brand’s support page is among the returned results: the brand’s site is a source.
The competitor is named with an advantage: it is recorded as the competitor’s presence.
A third-party forum among the results: it enters the external sources with its domain.
Search count
The number of searches is the one Perplexity declares in the answer. If it does not declare it, the returned search items are counted; if there are none, the number stays “not reported” and the receipt marks the cost as uncertain.
The sample receipt of a Perplexity answer about a fictional brand: model, tokens, declared searches, market and outcome; there is no service tier, because Perplexity does not report it.
Frequently asked questions
What changes between the API and Perplexity’s website, why there is no reading without search, and how Perplexity is compared with the other assistants.
Not necessarily: the website may answer differently, for example because of account settings. TrustInsider measures Perplexity’s official API and writes so next to every measure.
Because it answers only after searching. A “from memory” reading of Perplexity would be data that does not exist: for Perplexity only the search cell is read.
Yes, on the same questions, in the same market and only with search on for both. The comparison shows who names the brand and which pages it uses, not which assistant is better.
The answer stays valid if it is complete, but the number of searches is marked “not reported” and the cost stays uncertain in the receipt: no number is assumed.
Related paths
Perplexity shows the search results each answer is built on. Alongside you find AI Visibility with every assistant, the path to using the platform, the methodology we sample with and the answers from support.
See when ChatGPT, Gemini and the other assistants name, recommend or cite your brand: the answers, question by question.
ReadLearn how to use TrustInsider: choose the brand and the questions, read the collected answers, update your content and compare the results.
ReadHow every TrustInsider measure is built: sampling of AI answers, sources and rights, index calculation and declared limits.
ReadHelp with access, credits, collections, integrations and data: guides, FAQs, the site assistant and support requests from your workspace.
Read