This entry is compiled from primary sources: Perplexity AI's own documentation, release notes, and licensing or pricing pages.
It has not been hands-on tested by us. Where a limit, a price or a requirement is stated below, it is the published figure at the time of writing rather than a measured result.
Anything we have not verified is left out instead of guessed at.
Perplexity AI Review (2026) — The Free Tier Is a Citation Engine, Not a Chatbot
FreePerplexity
Perplexity gets filed under chatbots because it has a chat box, and that filing does it a disservice. It is an answer engine: you ask a question, it searches, and it returns a short synthesised answer with numbered links to the pages the claims came from. The links are the product. Everything else is packaging.
That single design decision is why it is worth having open in a tab even if you already pay for a general assistant, and it is available on the free plan.
Why the citations change how you use it
The failure mode of every general chatbot is the confident wrong answer, and the reason it is dangerous is that a fluent paragraph gives you nothing to check. Perplexity's output is shaped so that checking is one click, which quietly changes your behaviour: you stop asking it to be right and start using it to find out who says what.
It is at its best on questions with a documented answer somewhere — current pricing, a specification, a policy change, what happened in a news story last week, which of three products supports a given format. It is at its worst on questions where the honest answer is that nobody has written it down, because the shape of the interface still produces a tidy sourced-looking paragraph.
The citations also make its errors legible rather than invisible. When it misreads a source, you can see the source. That is a meaningfully different relationship with a tool than trusting an unsourced summary.
What the free plan includes, and the number nobody publishes
Free use gets you the core loop: search, synthesise, cite, follow up. Basic searches are effectively unmetered for ordinary use. The deeper Pro Search mode, which runs a multi-step research process instead of a single pass, is the one that is rationed on the free plan.
Perplexity does not publish the free allowance anywhere in its documentation, and it has changed it more than once. This is a criticism, not a footnote: you cannot build a habit around a quota you are not told and that can move without notice. Every specific daily figure circulating online is somebody's observation on a particular day rather than a commitment, and we are not going to repeat one here as though it were a fact.
What Pro is actually for
Perplexity's own help documentation is clearer about the paid tier than the free one. Pro is described as delivering roughly ten times as many citations per answer, extended access to Pro Search, a choice of frontier models from several vendors rather than only the house model, higher file and image upload allowances, and its Research and Create features.
Read that list again with an eye on what is missing from it: the citation behaviour itself. The thing that makes Perplexity distinctive is not paywalled. You are paying for depth and volume, which is a defensible upsell and a rarer one than it should be.
Pros
- Every answer arrives with sources you can open and check.
- The distinguishing feature is on the free plan, not held back for Pro.
- Strong on current, documented questions where a general chatbot's training data has gone stale.
- Errors are visible rather than hidden, because you can inspect what it read.
Cons
- The free quota is undocumented and has changed repeatedly.
- Rigid compared with a general assistant for drafting, brainstorming or code.
- Confidently synthesises even when the underlying sources are thin or contradictory.
- Answer quality is bounded by what is indexable, so it inherits the open web's blind spots.
- Pro's model line-up shifts as vendor deals change, which makes long-term comparison awkward.
Best alternatives
- You.com — the nearest competitor built on the same search-plus-citation premise (chatbots).
- ChatGPT — better when you want a collaborator to draft with rather than an answer to verify (chatbots).
- Gemini — free with a Google account and increasingly wired into Search itself (chatbots).
FAQ
How many free searches does Perplexity give you per day? Perplexity does not publish a number, and the allowance has been changed more than once. Basic searches are effectively unmetered for normal use; the deeper Pro Search mode is the metered one. Any specific daily figure you find quoted online is somebody's observation on a particular day, not a documented commitment, so treat it accordingly.
What does Perplexity Pro actually add? Perplexity's own help documentation describes Pro as giving roughly ten times as many citations per answer, extended access to Pro Search for multi-step research, a choice of underlying frontier models rather than the house one, more file and image uploads, and access to its Research and Create features. The core citation behaviour that makes Perplexity worth using is on the free plan.
Is Perplexity better than a normal chatbot? For anything you need to verify, yes, because it is built to show its sources and a general chatbot is not. For open-ended drafting, brainstorming or coding, no, because the search-and-summarise shape that makes it trustworthy also makes it comparatively rigid. Most people end up using it alongside a general assistant rather than instead of one.