How to Use Perplexity AI for Research: A Step-by-Step Tutorial

Perplexity AI is the research tool that answers with sources attached: ask a question, and instead of a list of links you get a written answer with numbered citations you can click through to verify. Founded in 2022 by former OpenAI researcher Aravind Srinivas, it passed 100 million monthly users by mid-2026. This tutorial covers every feature that matters, from your first question to a full Deep Research report.

The short answer: To use Perplexity AI for research, sign up free at perplexity.ai, type a specific question, read the cited answer, then refine it with follow-up questions. Use Focus modes to narrow sources, run Deep Research for long reports, organize work in Spaces, upload files to ask questions about them, and pick a model on Pro. Perplexity Pro costs $20 a month ($200 a year); the free plan gives unlimited basic searches and around 3–5 Pro Searches a day.

Step 1: Sign Up and Set Up Your Account

Go to perplexity.ai and create a free account with your email or Google account — no credit card is required. Signing in unlocks your search history, saved threads, Spaces, and follow-up suggestions. Students should check Education Pro at $10 a month — the full Pro feature set at half price for verified students and educators.

Step 2: Ask Your First Question — and Read the Citations

Type your question as a full sentence, not a keyword string. "What is the current market size of AI project-management software, and which companies lead it?" gets a far better answer than "AI project management market size." Perplexity searches the live web and writes an answer with numbered citations linking to the exact source page.

The most important habit: click the citation numbers. Perplexity's answer is only as good as the sources behind it, and it can occasionally misread a source or blend two figures. Clicking through takes seconds and is the difference between research and rumor. Our best AI tools for students guide covers more tools that pair well with Perplexity.

Step 3: Use Focus Modes to Narrow Your Sources

By default Perplexity searches the open web. The Focus selector next to the search box narrows the search to one kind of source — the single biggest quality upgrade on the free plan. Web is the default: the whole open web, best for general questions and current events. Academic restricts results to peer-reviewed papers and scholarly sources — the right choice for essays, theses, and literature reviews. Social pulls from Reddit and similar communities, the fastest way to find real-user opinions rather than marketing copy. Finance gives real-time stock data, earnings, and market news. Video retrieves relevant YouTube content, and Writing drafts without web retrieval.

Step 4: Build a Thread With Follow-Up Questions

Perplexity keeps each question and its follow-ups in a thread, and the AI remembers the context. After your first answer, ask a follow-up in the same thread — "Now compare the top three vendors on price" — instead of starting a new search. Perplexity also suggests follow-ups under each answer.

Step 5: Run Deep Research for Long Reports

When a quick answer is not enough, Deep Research is Perplexity's most powerful mode. It breaks your question into subtasks, runs dozens of parallel searches across hundreds of sources, cross-references the data, flags contradictions, and writes a structured report with sections and citations. A typical query takes 2–5 minutes, and since a July 2026 update it can also generate presentations, spreadsheets, and dashboards from the research. Deep Research is available to Pro subscribers, with a limited number of tries per day on the free plan. Save it for questions that deserve a report: a market analysis, a literature review, a competitive comparison. Use Pro Search for everything in between.

Step 6: Organize With Spaces (Projects)

Spaces — now also called Projects — are persistent workspaces where you organize research threads by topic or project, with custom AI instructions and file uploads. Create one Space per project — a thesis chapter, a market report, a client brief — upload your key documents, set an instruction like "Always cite academic sources and flag any claim without a source," and every thread inside that Space follows it. Free users get a small number of Spaces with limited uploads; Pro subscribers get more Spaces and up to 50 files per Space.

Step 7: Upload Files and Ask Questions About Them

Perplexity can read the documents you upload — PDFs, images, and other files — and answer questions about them alongside web sources. Upload a research paper and ask "What methodology did this study use, and what were its limitations?" or upload an earnings report and ask for a summary of the revenue figures. Pro subscribers get unlimited file and image uploads. The answer is grounded in your document and the live web at the same time.

Step 8: Pick the Right Model on Pro

Pro subscribers can choose which AI model answers each question: GPT-5, Claude, Gemini, Grok, and Perplexity's own Sonar models, with newer versions such as GPT-5.2, Claude Sonnet 4.6, and Gemini 3.1 Pro appearing as they are released. The practical rule: use Sonar for fast factual questions, Claude for long nuanced writing, and GPT-5 for structured analysis. Pro also adds image generation, a $5-a-month credit toward the Sonar API, and priority access during busy periods. A $200-a-month Max tier adds Model Council and the Comet AI browser.

Worked Example: Researching a Market in 20 Minutes

Here is the full workflow on a realistic question — "Is the AI project-management software market a good space to enter in 2026?" Start in Web mode and note the headline figures: the market was valued at $4.28 billion in 2026, projected to reach $8.9 billion by 2030 at a 20.1% compound annual growth rate. Click the citations and confirm the numbers come from a real market research firm.

Next, switch to Finance mode and ask which public companies lead the space and how their revenue is trending. Then run a Deep Research query: "Compare the top five AI project-management vendors on pricing, AI features, and target customers, with sources." While it runs, create a Space called "AI PM Market" for your saved pricing pages. When the report lands, check it against the cited sources — especially any market-share claim — and ask a follow-up: "What are the main risks or criticisms of these growth forecasts?" In about 20 minutes you have a sourced, organized brief instead of a pile of open tabs.

Read Also: ChatGPT vs Claude — how Perplexity's research answers compare with the two big chatbots. Teachers should also read Best AI Tools for Teachers.

Common Mistakes to Avoid

Three mistakes undo most Perplexity research. First, copying a figure without clicking its citation — AI answers can misread a source, combine two different numbers, or cite a page that says the opposite of the summary. Second, asking vague questions: "Tell me about AI" returns a shallow overview, while a specific question with a date and a scope returns something usable. Third, using Deep Research for a question a single search could answer — it takes minutes and spends your limited quota, so save it for reports, not quick facts.

Frequently Asked Questions

Is Perplexity AI free?

Yes. The free plan gives unlimited basic searches with full citations and around 3–5 Pro Searches a day, no credit card required. Pro costs $20 a month ($200 a year) and adds unlimited Pro Search, all AI models, Deep Research, and unlimited file uploads.

What is Perplexity Deep Research?

Deep Research is Perplexity's agentic mode: it breaks a question into subtasks, runs dozens of parallel searches across hundreds of sources, and writes a structured, cited report in 2–5 minutes. Pro subscribers get it, with limited tries per day on the free plan.

Which is better for research, Perplexity or ChatGPT?

Perplexity is better when the answer must be current and sourced, because it searches the live web and cites every claim. ChatGPT is better for brainstorming and drafting. Many use both: Perplexity to find and verify facts, ChatGPT to write them up.

Is Perplexity AI accurate?

Usually, but not always. Perplexity grounds its answers in live sources and shows its citations, which makes errors easier to catch — but it can still misread a source or blend figures. Always click through to the cited source before quoting a number or claim.

Related: Best AI Tools for Students | ChatGPT vs Claude | How to Use ChatGPT to Write a Resume