How to Use Perplexity AI: The Beginner's Guide (Ask, Verify, Then Go Deeper)
CodexFuse Team · 2026-10-28T02:18:01.028Z · 14 min read
Learn how to use Perplexity AI: ask full questions, pick the right mode, check the citations, and use Pro Search, Research, and Spaces — free, no account needed to start.
To how to use Perplexity AI well, open perplexity.ai or the app, type a full question in plain language, and press Enter. Perplexity searches the live web, writes a short answer, and attaches numbered citations to the sources behind every claim. Then click two or three of those citations before you trust anything important, and ask follow-up questions in the same thread when the answer needs to go deeper. That loop — ask, verify, follow up — is the whole skill, and it takes about a minute to learn. This guide walks through it step by step, including the search modes, the free-versus-Pro line, and the mistakes that make beginners get shallow answers.
Perplexity is an answer engine, not a chatbot and not a search engine. A search engine hands you ten blue links and leaves the reading to you. A chatbot writes a confident reply but often cannot show you where its facts came from. Perplexity does both jobs at once: it searches the web in real time, then writes the answer with clickable sources attached. That makes it the right tool when fresh, sourced information matters — checking a price, researching a topic, or verifying a claim before you repeat it. If you want help writing or thinking instead, use ChatGPT, which is built for conversation rather than search.
How do you use Perplexity AI?
Perplexity AI is an answer engine that searches the live web and writes a cited answer to a question asked in plain language. You use it by typing a complete question, reading the answer, opening the numbered citations to check the claims that matter, and asking follow-up questions in the same thread until the answer is solid enough to act on.
- Ask one clear question. Write it the way you would ask a knowledgeable friend, with the context that decides a good answer.
- Pick the right mode. Best for quick facts, Pro Search for comparisons and trade-offs, Research for a full report, Academic for papers.
- Read the answer, then check the citations. Open at least two sources for anything you will quote, publish, or spend money on.
- Ask follow-ups in the same thread. Narrow, compare, and challenge the answer instead of starting over.
- Demand better sources when the answer is weak. Tell it to use only primary sources and list what is still uncertain.
- Save repeatable work in Spaces. Group threads by project so the context carries over.
- Upload files when the question is about your own documents. Ask about a PDF or CSV and search the web in the same thread.
What is Perplexity AI?
Perplexity AI is a free answer engine that launched in 2022. You ask a question in normal language, and instead of a list of links it writes a short, direct answer with numbered citations so you can check where each part came from. It works in any browser at perplexity.ai, and it has free apps for iPhone and Android. You can even start without making an account, though an account gives you search history and limited file uploads.
The difference from Google and ChatGPT is simple. A normal search engine points you to pages and leaves the reading to you. A standard chatbot writes an answer, but the free versions often pull from training data and do not always show their sources, which is one reason they sometimes get facts wrong. Perplexity searches the web in real time, then writes the answer with clickable sources attached. That makes it a strong choice when fresh, sourced information matters, and a weak choice when you want long-form drafting or creative writing — for that, a broader assistant fits better.
One warning belongs up front, and Perplexity's own help pages carry it: responses are generated by AI and may contain mistakes. A citation shows where information came from, but it does not guarantee Perplexity interpreted the source correctly. The citations are only valuable if you actually open them.
How do you ask a good question in Perplexity?
Ask a full question, not keywords. That is the single biggest difference between Perplexity and Google, and the mistake most beginners make. Typing "podcast mic" gives Perplexity almost nothing to work with. Typing "Best budget mic for podcasting under $100" gives it the product, the use, and the budget in one line — and the answer is better for it.
A strong prompt includes four things: the decision you are trying to make, the constraints, what "good" looks like, and what to cite. A market-research version looks like this: "I am researching budget podcast microphones in the US. Give me a concise overview of the top 5 options under $100, with the price and the main trade-off for each. Include cited sources for every price, and list anything you are uncertain about." That prompt naturally pushes for data, cited sources, and factual accuracy rather than a generic summary.
Two more habits help. Say who you are and what the answer is for — a business owner in Leeds gets a different answer from a student, because the sources and the framing change. And ask for a format: a table, a short list, or a summary under 150 words. Perplexity follows format instructions well, and a table makes a comparison far easier to check than a paragraph. If you want to get sharper at this part, the skill transfers to every AI tool — see our guide on how to write better AI prompts.
| Weak prompt | Strong prompt | Why it works better |
|---|---|---|
| podcast mic | Best budget mic for podcasting under $100 for a Mac | Product, use, budget, and platform in one question |
| best laptop | Best laptop under $800 for a college student who edits video | Budget, user, and workload decide the answer |
| is solar worth it | Is rooftop solar worth it for a 2,000 sq ft home in Texas? Include payback time and cite sources | Location, size, and the number that matters |
| what is inflation | What is the current US inflation rate, and what drove it this year? Cite official sources | Asks for the current figure and the cause, with sources named |
Which Perplexity mode should you use?
Perplexity has four search modes, and choosing the right one is the biggest difference between a shallow answer and a useful one. Start with the default and step up only when the answer is thin — that keeps your Pro Search allowance for the questions that actually need it.
| Mode | What it does | Use it for |
|---|---|---|
| Best | The default. Picks the most suitable model for your question, with no quota limits | Quick facts and everyday questions |
| Pro Search | Draws on a wider set of high-quality sources and lets you choose an AI model | Comparisons, analysis, and anything where depth matters |
| Reasoning | Uses thinking models that break a complex question into multiple search steps | Tricky, multi-part problems |
| Research | Runs dozens of searches, reads hundreds of sources, and writes a full report | In-depth reports and background research |
The rule of thumb is simple. If your question has the word "best," "compare," "strategy," or "why," you probably want Pro Search. If you need a report — a literature review, a market brief, a background dossier — use Research, which delivers its reports in two to four minutes. Research automatically selects models for you, so you cannot manually pick a specific one inside that mode.
There is also an Academic focus, which targets scholarly sources: peer-reviewed papers, journal articles, and reputable academic publications. When you need academic work rather than blog opinions, flip to Academic and ask for the key points, the methods and sample size, the limitations, and where the field agrees or disagrees. That is how you keep "research" from becoming a pile of blog opinions.
How do you check a Perplexity answer before trusting it?
Open two or three citations for every important claim. That is the habit that makes Perplexity worth using, and skipping it is the habit that gets people into trouble. Every answer has numbered citations linking to the original sources, and Perplexity's own advice is to treat them as your safety net — especially for a number, a quote, or a price.
Source quality matters as much as source count. Prefer reputable sources: official documentation, respected outlets, and primary research. If the citations are weak, ask Perplexity to fix that in the thread: "Use only reputable sources — official docs, major publications, or primary research. Avoid SEO blogs. Show sources for key points." Then click the citations again. This is the difference between AI-generated content that looks confident and research you can stand behind.
For anything you will publish, use the thread to force precision before you copy a word. Ask "Which claim is most uncertain, and why?" Ask "What would change your conclusion?" Ask "Give me the 3 strongest sources and summarize each." And if sources disagree, ask it to compare the two opposing viewpoints and cite both. A fact-checking prompt that works well is: "Verify these claims: [paste claims]. For each, say True, False, or Unclear, provide supporting citations, and explain the reasoning."
What is Pro Search in Perplexity?
Pro Search is Perplexity's advanced mode for nuanced questions and complex topics. It goes beyond basic retrieval and synthesizes across a broader, higher-quality set of sources, and it lets you choose which AI model does the work. Use it for market analysis, competitive analysis, buying decisions with trade-offs, and anything with multiple constraints.
Free users get a limited number of Pro Searches per day; Pro extends that allowance. That limit is the practical reason to start every question in the default Best mode and step up only when the answer is thin. If you burn Pro Searches on questions Best could have answered, you will run out on the one that actually needed the deeper pass.
A good Pro Search prompt names the comparison and the evidence it must carry. For example: "Compare Competitor A vs Competitor B for small-business invoicing. Include pricing, target market, positioning, and key differentiators. Cite source material for each claim, and flag anything uncertain." The model choice matters here too — Pro lets you pick from advanced models including GPT-5.2, Claude Sonnet 4.6, and Gemini 3.1 Pro, according to Perplexity's help centre.
What is Research mode in Perplexity?
Research mode is Perplexity's "do the heavy lifting for me" feature. It conducts in-depth research and analysis for you: dozens of searches, hundreds of sources, and a structured report at the end, delivered in two to four minutes. Use it for academic research, literature reviews, long-form briefs, and complex questions with lots of moving parts.
A literature-review prompt shows the format: "In Academic focus, find 8 to 12 academic publications from the last 5 years on [topic]. Summarize key points, methods, and limitations. Prioritize peer-reviewed journals. Output as: citation, findings, caveats." The structure matters because a Research report is long — if you do not tell it how to organize the output, you will spend your time reorganizing instead of reading.
Research is also the mode where the verification discipline pays off most. A report that cites hundreds of sources still needs a human to open the three or four citations behind its central claims. Treat the report as a first draft of your research, never as the finished product, and never paste a Research report into your own work without checking the sources and rewriting it in your own words.
How do you use Spaces and Projects in Perplexity?
Spaces let you organize work, set custom instructions, and collaborate with others as research partners. Think of a Space as "my templates and rules for this project." Examples: weekly competitor monitoring, customer discovery and user research, content creation that matches your brand tone, or lesson plans and learning modules.
The practical setup is one Space per client or per big piece of work. You add prompts and files to the whole Space, so every search inside it starts with the same context, and you can invite members to collaborate. That keeps the research tidy and saves you re-explaining the background every time. The newer Projects view works the same way: sessions and files grouped by topic, with the same custom instructions applied across them.
Privacy is straightforward: sessions and Projects are private by default. A session is only shared if you select Share, and a Project is shared by inviting members. If you are researching something sensitive — a health question, a legal matter, a business negotiation — keep it in a private Space and do not share the thread.
Can you upload files to Perplexity?
Yes. You can attach a file to a session, ask questions about it, and search the web in the same session. Perplexity lists PDFs, CSVs, audio, video, and images among the files you can upload. It is handy for comparing a supplier's proposal with what is on the market, or pulling the key points from a long report. Pro users get more file and photo uploads than free users.
The workflow that works best is to upload first, then ask in stages: "Summarize this report in five bullet points," then "Which of these claims can you verify against current sources?" then "What is missing from this report that a buyer would want to know?" That keeps the file and the web in conversation with each other instead of treating the upload as a dead document.
One caution: never upload a file containing passwords, API keys, or personal records to any online tool, including Perplexity. Config files, exports, and internal documents often hold secrets. Strip them out before uploading, or work with a local copy instead. The same rule applies to the research-assistant lane — if your work is mostly reading your own documents, NotebookLM is built around that job, answering only from the sources you upload.
Is Perplexity free? Free vs Pro
Yes, Perplexity is free to use. You can search for free, and free users get a limited number of Pro Searches per day. Perplexity Pro extends that allowance and adds advanced models, more uploads, and extended Research access. For most beginners, the free plan is the sensible place to start. Consider Pro once you regularly run into the free limits, or specifically need more research capacity, file analysis, or control over which supported AI model is used.
| Feature | Free | Pro |
|---|---|---|
| Price | $0 a month | $20 a month, or $17 a month when billed annually |
| Basic searches | Available, with limited daily usage | Available with higher limits |
| Pro Search | Limited number per day | Extended access |
| Model choice | Perplexity selects the model | Choose from advanced models, including GPT-5.2, Claude Sonnet 4.6, and Gemini 3.1 Pro |
| Citations per answer | Standard | Ten times as many citations per answer |
| File uploads | Limited | Higher upload and analysis limits |
| Research mode | Basic access | Extended access |
| Create files and apps | No | Yes — reports, spreadsheets, dashboards, and simple web apps from a prompt |
| Best for | Occasional research and quick answers | Frequent research, file analysis, and advanced features |
The prices above were checked on October 28, 2026: the $20 monthly figure from ByteTony's guide, verified September 2026, and the $17 annual-billing figure from Perplexity's own plan dialog the same day. Pricing, limits, and available models change regularly, so check the official Perplexity site before upgrading. There is also a Max tier at $167 a month when billed annually, aimed at unlimited usage and top performance, and an Education Pro tier for students and teachers. Pro subscribers also receive $5 in monthly API credits, credited on the first day of each month, if you want to use Perplexity's search API in your own tools.
Perplexity vs ChatGPT: which should you use?
Use Perplexity when the answer depends on what is out there right now, and an assistant like ChatGPT or Claude when you want help writing or thinking. Perplexity is built around search: it looks up the web in real time, and every answer includes numbered citations linking to its sources, so you can check each claim. ChatGPT is a broader general-purpose assistant that can also search the web and do research, but it covers a wider range of tasks beyond research — drafting, coding, brainstorming, and planning.
The practical split is simple. If you need a sourced answer — a price, a statistic, a comparison, a fact to publish — start in Perplexity. If you need help turning that research into something — a draft, a plan, a rewrite — take the verified facts to a chatbot. The two tools are complements, not rivals, and most people who research seriously end up using both. If you want the same beginner system for the other side, our guide on how to use Google Gemini covers the assistant that lives inside Google's own apps.
| Question | Start with Perplexity | Start with ChatGPT or Gemini |
|---|---|---|
| What does this cost right now? | Yes — live web search with cited prices | No — may quote an old price |
| Which of these two tools is better? | Yes — comparison with sources | Maybe — good for reasoning, weak on current facts |
| Is this statistic true? | Yes — verify against the cited source | No — it may repeat the claim without a source |
| Help me draft this email | No — not built for drafting | Yes — built for writing |
| Plan my week / brainstorm ideas | No — search-first workflow | Yes — conversation-first workflow |
What are the common mistakes beginners make with Perplexity?
Treating it like Google. Perplexity rewards natural questions with constraints, not keyword piles. If you type two words, you get a two-word answer. Write the question you actually want answered.
Trusting the first answer. Click the citations. Ask follow-ups. Use Academic mode for research. A confident-sounding answer with weak sources is still a weak answer, and the first draft is rarely the one you should act on.
Using the wrong mode. Pro Search for comparisons, Research mode for depth, Academic for papers. The default Best mode is fine for quick facts, but it will not do the deep work for you — and if you use Pro Search for everything, you will hit the free limit on the question that needed it.
Copying and pasting AI output as "done." Use Perplexity to accelerate thinking, drafts, and synthesis, but always review for accuracy and source quality. Anything you publish still needs human review, and the cited sources need opening. The citation is the start of verification, not the end of it.
Forgetting to save the work. A good thread is an asset. Save it in a Space, or it disappears into your history and you will redo the same research next month. Threads are also how Perplexity keeps context — a follow-up in the same thread is worth more than a fresh search.
Sources
- AIclicks, "How to Use Perplexity AI (Like a Pro) in 2026" — 7-step workflow, Web/Academic/Pro Search/Research modes, citation discipline, Pro $5 monthly API credits, subscription tiers. Updated August 22, 2026. aiclicks.io
- Lilach Bullock, "How to Use Perplexity AI: A Practical Beginner's Guide" — mode table (Best, Pro Search, Reasoning, Research), Research reports in 2–4 minutes, file uploads (PDFs, CSVs, audio, video, images), Projects, Pro features (10× citations, GPT-5.2, Claude Sonnet 4.6, Gemini 3.1 Pro). Updated October 24, 2026. lilachbullock.com
- ByteTony, "Perplexity AI for Beginners: Simple Guide (2026)" — full-question rule, citation check before trusting, Free vs Pro table ($0 vs $20 a month), where Perplexity falls short. Verified September 2026. bytetony.com
- Perplexity, "Select your plan" dialog on perplexity.ai — Pro $17 a month when billed annually, Max $167 a month when billed annually, Free "limited daily usage," Pro includes 4,000 bonus credits, deep research, top AI models, model choice, and creating polished documents and apps. Checked October 28, 2026. perplexity.ai
- BrightmindAI, "What Is Perplexity AI? A Beginner's Guide to AI Search" — launched in 2022, no account needed to start, answer engine vs search vs chatbot. Published June 30, 2026. brightmindai.medium.com
Frequently Asked Questions
Is Perplexity free to use?
Yes. You can search for free, and free users get a limited number of Pro Searches per day. Perplexity Pro extends Pro Search and adds advanced models, more uploads, and extended Research. Pro costs $20 a month, or $17 a month when billed annually, checked October 28, 2026.
How is Perplexity different from ChatGPT?
Perplexity is built around search. It looks up the web in real time, and every answer includes numbered citations linking to its sources, so you can check each claim. ChatGPT is a broader general-purpose assistant built for conversation, writing, and planning, and it can search the web too — but its answers do not always carry citations.
What is Pro Search in Perplexity?
Pro Search is Perplexity's advanced mode. It draws on a wider set of high-quality sources, links to each one, and lets you choose from several advanced AI models. Use it for comparisons, analysis, and anything where depth matters. Free users get a limited number of Pro Searches per day.
Can I upload documents to Perplexity?
Yes. You can upload files to a session and ask about them while also searching the web. Perplexity lists PDFs, CSVs, audio, video, and images. Pro users get more file and photo uploads. Never upload files containing passwords, API keys, or personal records.
Are my Perplexity searches private?
Perplexity says sessions and Projects are private by default. A session is only shared if you select Share, and Projects are shared by inviting members. Do not share a thread that contains sensitive personal or business information.