Perplexity

What is Perplexity?

Perplexity is an AI “answer engine” — a hybrid of a search engine and a chatbot. Ask a question in plain language and, instead of a page of blue links, you get a concise, written answer. What sets it apart is visible in every response: numbered inline citations that link back to the exact pages the answer was built from, so you can click a footnote and verify any claim yourself.

The product launched in late 2022 around a simple premise: a chatbot should show its work. Ordinary chatbots produce fluent text but can hallucinate, and there is no easy way to check them. Ordinary search hands you links but leaves the reading and synthesis to you. Perplexity sits between the two. It synthesizes an answer like a chatbot while grounding that answer in freshly retrieved, cited sources like a search engine.

Because it is conversational, the experience is iterative. You ask, you get a sourced answer, then you ask a follow-up and Perplexity keeps the context of the thread. It suggests related questions to help you go deeper, and depending on the “focus” you choose it can draw from the open web, academic papers, community discussions, or video. The result feels less like searching and more like having a well-read research assistant who always cites their references.

Key features

• Cited answers — numbered footnotes link to the source pages behind each statement.
• Live web search — answers reflect current information rather than a frozen training cutoff.
• Conversational threads — follow-up questions keep context, and suggested questions nudge you further.
• Focus modes — scope a query to the web, academic papers, social/community discussion, or video.
• Pro Search — a slower, multi-step mode that clarifies intent and searches more thoroughly before answering.
• Model choice (paid) — switch between leading frontier models from OpenAI, Anthropic, and Google, alongside Perplexity’s own Sonar models and dedicated reasoning models for harder problems.
• File uploads — ask questions about your own PDFs, spreadsheets, and images.
• Spaces / Collections — group threads by project, add custom instructions, and collaborate.
• Sharing — publish threads as public pages or share them with a link.
• Apps and API — iOS and Android apps with voice input, plus a developer API (Sonar) for building the answer engine into your own products.

Who it’s for / best use cases

Perplexity is at its best whenever the source of an answer matters as much as the answer itself.

• Students and researchers running literature reviews or fact-finding, who need every claim traceable to a real citation.
• Journalists, analysts, and knowledge workers who want a fast, sourced first pass on an unfamiliar topic before committing time to it.
• Comparison shoppers researching products, software, or travel, who want a synthesized summary of current information with links to dig in.
• Developers and technical users who need up-to-date documentation answers with references rather than a model’s stale recollection.

It shines on current events (“what’s the latest on…”), company and market research, quick technical lookups, and any question where checking the source is part of the job. It is comparatively weaker at long-form creative writing, open-ended brainstorming, and complex multi-file coding — tasks where a general-purpose chatbot is usually the better tool.

Pricing & plans

Perplexity offers a free tier that is genuinely useful rather than a crippled demo: unlimited standard searches and a capped number of the deeper Pro searches each day, which is plenty for casual research.

Pro costs $20/month (or $200/year) and is the plan most serious users land on. It raises the daily Pro-search limit substantially, unlocks model choice, and adds file uploads, image generation, and API credits. For anyone who reaches for Perplexity several times a day, the upgrade pays for itself in saved time.

Above Pro sit a higher-priced Max tier for power users who want the most generous access, and an Enterprise Pro plan for teams that need admin controls and security. Most individuals never need to go beyond the free tier or Pro.

Pros and cons in depth

The citation-first design is the real reason to choose Perplexity. Because every claim is footnoted, you can audit an answer in a single click — that turns the AI from a black box into a starting point you can trust and trace, which matters enormously for research, work, and anything you will act on. Pair that with live web access and you get answers that stay current, not a model guessing from months-old training data.

Model choice on Pro is a genuine advantage too: you can send a factual lookup to a fast model and a thorny reasoning problem to a dedicated reasoning model, without juggling separate subscriptions.

The trade-offs are honest ones. Answer quality is only as good as the pages Perplexity finds, so a thin or low-quality source set can produce a shallow or subtly wrong summary — spot-checking the citations is worth the few seconds it takes. The free tier’s cap on deep searches is fair but noticeable if you research heavily. And because it is tuned for grounded, factual answers, it is not the tool you want for long creative drafts or big coding tasks.

How it compares / alternatives

ChatGPT now has a search mode and is the stronger all-rounder for writing, coding, and brainstorming; pick it when you want one tool for everything and citations are secondary. Google’s Gemini (and AI-powered Google Search) is deeply tied into Google’s index and Workspace, and is a natural pick if you already live in Google’s ecosystem. Microsoft Copilot offers similar web-grounded answers wired into Windows and Microsoft 365. Choose Perplexity specifically when transparent, clickable sourcing is the priority and you want a focused research tool rather than a do-everything assistant.

FAQ

Is Perplexity free? Yes — the free tier gives you unlimited standard searches and a limited number of deeper Pro searches per day. Pro ($20/month) removes most of those limits and adds model choice and file uploads.

Does it actually cite its sources? Yes. Every answer includes numbered citations that link to the pages it used, and you should click through on anything important, because summaries can occasionally miss nuance.

How is it different from ChatGPT? Perplexity is search-first and always grounds answers in live, cited web sources. ChatGPT is a broader assistant that is stronger at writing and coding but historically less focused on transparent sourcing.

Which AI model does Perplexity use? The free tier uses Perplexity’s own optimized models; Pro subscribers can choose among leading models from OpenAI, Anthropic, and Google as well as Perplexity’s Sonar models, and switch depending on the task.

Verdict

Perplexity is the best mainstream tool for answers you can trust and trace. Its citation-first, live-web approach solves the single biggest weakness of chatbots — you no longer have to take the AI’s word for it. It is not the pick for creative writing or heavy coding, and you should still glance at the sources on anything important. But for research, fact-checking, and staying current, it earns a permanent place in your toolkit. Casual users are well served by the free tier; anyone who researches daily should buy Pro.