Module 422 min read · Mastering Perplexity

Focus Modes & Pro Features

Focus modes are Perplexity's most underused capability. By default, Perplexity searches the entire web — which is usually what you want. But when you know the type of source that will best answer your question, restricting search to that source type dramatically improves result quality. This module covers every Focus mode and the Pro features that unlock deeper research.

Focus modes — restricting your search source

Focus modes appear as icons below the search bar. Selecting one tells Perplexity to search only within that domain of content rather than the full web. The difference in output quality for specialized queries is significant.

🌐

Web (default)

Searches the full web. The right default for most queries — current news, general information, product research, factual questions. Use this unless you have a specific reason to narrow the source domain.

Best forGeneral research, current events, product comparisons, factual questions about publicly known topics

🎓

Academic Pro

Searches academic databases — PubMed, arXiv, Semantic Scholar, and similar peer-reviewed sources. Results are limited to published research and scholarly work. This is the mode that transforms Perplexity into a legitimate research tool for students and researchers — getting peer-reviewed sources rather than general web content.

Best forLiterature reviews, fact-checking scientific claims, finding studies on specific health or science topics, academic research projects, understanding the evidence base for any claim

💻

Code

Searches Stack Overflow, GitHub, technical documentation, and developer-focused content. Optimizes for code examples, error solutions, library documentation, and technical explanations. The synthesized answers lean toward concrete, implementable solutions rather than conceptual overviews.

Best forDebugging errors, finding implementation examples, understanding library APIs, looking up syntax, finding solutions to specific technical problems

▶️

YouTube

Searches YouTube specifically and returns relevant videos with timestamps and summaries. Rather than just finding videos, it synthesizes what those videos contain — letting you understand video content without watching. Particularly useful for tutorials, reviews, and how-to content.

Best forFinding tutorials on specific topics, comparing product review videos, understanding what's been covered on YouTube about a topic, finding visual explanations

💬

Searches Reddit discussions specifically. This is uniquely valuable for real user opinions, experiences, and community knowledge that doesn't appear in polished articles. What do actual users think of X product? What problems do people actually encounter with Y? Reddit answers these where official sources don't.

Best forReal user opinions on products, community knowledge, personal experience with services, understanding what problems people actually encounter, niche hobby and interest topics

Pro features worth knowing

Model selection

Perplexity Pro lets you choose the underlying language model used for synthesis. Options typically include GPT-4o, Claude Sonnet, Gemini Pro, and Perplexity's own models. The search layer is the same regardless — the model choice affects synthesis quality, writing style, and reasoning depth.

When to switch models: Claude Sonnet tends to produce better-written, more nuanced synthesis. GPT-4o tends toward comprehensive coverage. For most research queries, the default Perplexity model is fine — switch when you need a specific writing style or deeper reasoning on a complex topic.

File upload and analysis

Pro users can upload documents (PDFs, Word files, spreadsheets) and ask Perplexity to analyze them — often combined with web research. Upload a report and ask Perplexity to compare its claims against current data. Upload a contract and ask it to flag unusual clauses relative to industry standards. The combination of your document plus live web context is powerful.

Image generation

Pro includes basic image generation via DALL-E or Stable Diffusion. This is a convenience feature rather than a serious image generation capability — for professional image work, ChatGPT's DALL-E 3 integration is stronger. But having it available within Perplexity means you can research a topic and generate related visuals without switching tools.

The Academic focus is a game-changer for students

If you're in school and need to find peer-reviewed sources, Academic focus mode does in seconds what used to take hours in Google Scholar. Ask a research question, switch to Academic, and get synthesized answers from actual published studies with full citations. The citations link directly to the papers. This is the most academically legitimate AI research tool available.

The Focus mode decision rule

Ask yourself: "What type of source would give me the best answer to this specific question?" If it's peer-reviewed research — Academic. If it's real user experience — Reddit. If it's code examples — Code. If it's video tutorials — YouTube. If you're not sure — Web. Matching your query to the right source domain is the difference between a mediocre answer and a great one.