AI Tools / research

Perplexity: The AI Search Engine That Cites Its Work

★★★★★ 4.8 (12000)

Real-time AI-powered research with inline citations. The closest thing to a research analyst that fits in a browser tab.

Visit Perplexity →
Free + paid plans

Perplexity is what Google would be if it answered questions instead of throwing ten blue links.

AI answer engine open for a research query, clean desk, search and chat interface, Perplexity tool review
Photo by Planet Volumes on Unsplash

Why citations change everything

Every Perplexity answer comes with numbered footnotes that link to the actual sources. This solves the #1 problem with LLM output: you can’t trust it. With citations, you can verify in two clicks. That alone makes Perplexity the right choice for any work that will be quoted, published, or used to make a decision.

Researching sources on a MacBook, developer desk, screen with results
Photo by Christopher Gower on Unsplash

Where Perplexity beats ChatGPT

Anything that requires current information: news, prices, recent research, this quarter’s industry numbers. ChatGPT and Claude have training cutoffs and only browse when explicitly told to. Perplexity searches first, always.

Pro tier worth it?

Yes, if you do research more than once a week. Pro unlocks better models (Claude, GPT-5), longer reasoning, file uploads, and significantly more searches per day. It also unlocks Spaces — persistent research collections you can come back to.

Researcher noting findings by hand, warm desk, notebook and keyboard
Photo by Kelly Sikkema on Unsplash

How to use it well

Treat Perplexity less like a chatbot and more like an analyst. Ask one question. Read the citations. Ask the follow-up. Keep going until you have the truth. Five turns of Perplexity beats an hour of regular Googling for most research tasks.

Key features

  • Real-time web search
  • Inline citations
  • Focus modes (Academic, Reddit, YouTube, etc.)
  • File uploads
  • Spaces (collections)
  • Pro Search with deeper reasoning

Pros & cons

Pros

  • Live web data with inline citations
  • Multiple model backends (GPT, Claude, Sonar)
  • Follow-up question chaining
  • Built-in academic and finance focus modes

Cons

  • Free tier limited on advanced models
  • Citations sometimes need verification
  • Less conversational than ChatGPT/Claude

Best for

  • Research
  • Fact-checking
  • Current events
  • Competitive analysis
  • Citation-grade work