What it actually is
Perplexity is an answer engine rather than a chatbot. You ask a question and it returns a short synthesised answer with numbered citations linking to the pages it drew on. The difference from a chatbot is accountability: you can see where each statement came from and judge it yourself.
What it’s genuinely good at
- Citations by default. This changes the workflow from “trust and publish” to “check in five seconds”. For anything factual, that is the whole ballgame.
- Getting oriented fast. New topic, unfamiliar jargon, need the landscape in two minutes — it collapses what would be twenty open tabs into one answer with links.
- Follow-up questions that keep context. You can drill down without restating the topic, which is where it beats a conventional search engine.
- Focus modes. Restricting a search to academic papers, videos, or discussion threads is genuinely useful and easy to forget about.
Where it falls short
Citations can be loose. A source link sometimes supports the general topic but not the specific sentence it is attached to. It is far better than no citation, and still not the same as a verified reference — check any figure you intend to repeat.
Reasoning is not its strength. For analysis, judgement or anything requiring nuance, a model like Claude does noticeably better. Perplexity retrieves and summarises; it does not think hard.
It is not a writing tool. The output is reference prose — serviceable, not publishable.
Limited free searches. The free tier allows a modest number of deeper searches per day before nudging you to subscribe or fall back to a quicker mode.
Who should use it
Anyone doing research, competitive analysis, technical lookups, or trying to understand a field quickly. It has genuinely replaced a large share of ordinary Googling for us.
Choose something else if: you need careful reasoning or long-form writing (Claude), you want sources inside a longer conversation (ChatGPT with browsing), or you are checking scientific evidence specifically (Consensus, which shows whether papers agree or contradict).
Pricing
Free access covers casual use with a limited number of deep searches daily. Pro, around $20 a month, raises limits substantially and lets you choose stronger underlying models. For anyone who researches as part of their job it pays for itself quickly; for occasional use free is enough.
Alternatives worth considering
- Consensus — purpose-built for scientific questions, showing which papers support or contradict a claim.
- ChatGPT with browsing — better when research is mixed with drafting or coding in one session.
- You.com — a comparable answer engine with a stronger privacy stance and a smaller index.
Our verdict: the single most useful research tool to appear in years, and the default we now open instead of a search box. Treat citations as leads rather than proof, use it to get oriented, and verify anything you plan to publish or bill for.
Frequently asked questions
Is Perplexity free?
Yes, with a daily limit on deeper searches. Pro removes the caps and unlocks stronger models.
Are the citations reliable?
Mostly — but not perfectly. A linked source sometimes supports the topic rather than the exact claim. Check the source before quoting a specific number.
Perplexity or Google?
Google when you know the exact page you want. Perplexity when you need an answer assembled from several sources, or when you are exploring something unfamiliar.
Perplexity or ChatGPT?
Perplexity for research with sources. ChatGPT for drafting, coding and mixed tasks where citations matter less.