Perplexity Pricing and How to Use It: Which Orbit to Start From

What differs by orbit

When to use Perplexity, and when to use a neighboring star

A 10-minute test route in the free orbit

Frequently asked questions

A closer look

What kind of star is Perplexity, and what is it used for?

It's a search-based AI that attaches sources to every answer. When you enter a question, it searches the web in real time, picks the links it finds, and a language model summarizes them. This approach is called RAG, a technique that builds answers from retrieved material. According to the Namuwiki description, it has an edge for questions that need up-to-date information and for the problem of AI making things up (hallucination). In return, the model's own versatility may fall short of other conversational AIs.

One thing to know: it doesn't rely only on its own models, and tends strongly toward drawing on other companies' models such as GPT, Gemini, and Claude. Its App Store description also says it is based on models from OpenAI, Anthropic, Meta, and others.

The most useful feature is source citation. Supporting sites are attached at the end of paragraphs or sentences, so you can click a link to check important figures right away. For example, if you ask about "latest iPhone specs," you get a summary based on official information and news sites, along with links. You can also narrow the search scope. You pick a source type, such as academic papers only, mainly major news sites, or including YouTube, which suits research for reports or checking market trends. It also offers a summary that shortens a long article to a few lines, iOS and Android apps, and an extension for searching straight from the address bar. In the same company's Comet browser, Perplexity is the default search engine, so you don't need to set anything up.

How far does free go, and how much is paid?

Regular search stays free. Namuwiki says even free users can search without limits. What gets blocked is Pro search. It greatly expands the sources and length of the answer, and a free account gets at most 3 per day. Once you've used them up, prompts to upgrade to Pro show up regularly. Keep that in mind as you use it.

Paid Pro is listed at about $20/month in one external review. The Korean App Store shows Perplexity Pro at KRW 33,000 as an in-app purchase, but the billing cycle isn't in the excerpt of that screen. The two amounts are hard to compare directly because they were checked at different times and come through different payment routes. Namuwiki says that on Pro you can choose third-party models such as GPT, Claude Sonnet, and Grok, and that some models require the higher MAX plan. Enterprise Pro, for companies and research institutions, is said to be a custom quote based on usage scale.

To sum up: try the free version for a few days first, and when you feel 3 Pro searches aren't enough, check the limits on the official pricing page and decide. The limits have changed within the past few months, so it's safer to check once more right before you pay.

What are the downsides?

The biggest controversy is the change in limits. Namuwiki says that from December 2025 the free trial began requiring a payment method to be registered, and that around February 2026 the Pro Deep Research limit (a feature that investigates in depth and answers like a report) was cut from 600 per day to 20 per month. It also records user backlash that there was no advance notice. However, a wiki is text that users edit, so treat this as "there are claims like this" rather than established fact.

There are weaknesses on the model side too. The default Sonar model is fast and good for search, but it is rated weak at creative or theoretical answers. Attached files such as images are rejected at upload if judged inappropriate. Files with even slight exposure or suggestions of sexual content are reportedly blocked as well.

App Store reviews include complaints about frequent errors and slow customer support replies. There are also posts saying it sometimes states wrong things without references. These are individual reviews and don't apply to everyone. Still, if you use it for tight-deadline work like a thesis, keep a backup tool.

Is what I type, and my personal information, safe?

Look carefully at the default settings. According to Proton's analysis, unless you use Incognito mode, on the Free, Pro, and Max personal plans your search data may be used to improve AI models. You can turn this off in account settings, but turning it off can't delete data already collected, and it only applies to data collected from then on.

Incognito mode is easy to misunderstand. The official privacy policy says that even in Incognito mode, third parties such as the company or the websites you visit can still track you. It only keeps other people using the same device from seeing your earlier history. The policy also says sensitive information, such as health details in the prompts you enter, may be collected. Also remember that it is not an end-to-end encrypted service, so the company can access the data if needed.

A class action lawsuit was filed in 2026 alleging that Incognito search data was shared for advertising purposes, but Proton said this claim has not yet been proven. Either way, the basic rule is not to enter national ID numbers, health information, or company secrets. If you handle confidential internal documents, check the terms of the enterprise plan before the personal one.

Who is it for, and how do I get started?

It suits people who want to check sources themselves. Typical users are students and office workers gathering material for reports or blog posts, and people who often check news and product updates. It's also good for skimming "recent trends in this industry" before a meeting. On the other hand, if your main use is creative work like fiction or ad copy, or long theoretical explanations, a general-purpose conversational AI such as ChatGPT or Gemini may suit you better. Namuwiki also records that a growing number of users plan to move to those platforms in protest of the reduced limits.

Getting started is simple. Create an account and log in on the web or in the iOS or Android app. Then enter your question in the search box, adding a scope rather than asking vaguely. For example: "Summarize in 3 lines the effect of remote work on productivity, mainly using academic sources from the past year." When the answer appears, click the source links at the end of the sentences to check that the figures and dates are correct. Finally, open the AI data retention option in settings and decide whether to allow your data to be used for training.

This article is not a hands-on review. It is compiled from a wiki, the official privacy policy, an App Store description, and an external review. Pricing and limits keep changing, so check again right before you pay.