Recommended Free Tools
Perplexity launched Deep Research on February 14, 2025. It is an autonomous, multi-step web-research workflow that plans searches, reads sources, follows leads, and produces a cited report. Perplexity’s current Help Center calls the feature Research mode. Free accounts have limited access, while Pro accounts receive extended access; every report still needs human source checking.
What Perplexity launched
Deep Research is more than a longer answer to a normal search query. Perplexity describes it as a workflow that interprets a research question, plans an investigation, runs multiple searches, opens and evaluates source material, reasons over the findings, and composes a report with citations. Reports can be exported to a document or PDF and, where available, turned into a shareable Perplexity Page.
As an Amazon Associate I earn from qualifying purchases.
The original announcement appeared in Perplexity’s official blog archive on February 14, 2025: Perplexity’s blog archive. It followed OpenAI’s February 2, 2025 announcement of ChatGPT Deep Research, putting both products in the first wave of mainstream research agents.
Perplexity’s naming has since shifted. A February 2026 changelog still says “Deep Research,” while the Help Center article updated July 16, 2026 uses “Research mode.” The two labels can therefore appear in different parts of the product.
#1 Best Overall
How Research mode works
- Interprets the brief: It identifies the question, constraints and likely subtopics.
- Plans searches: It breaks a broad assignment into multiple queries instead of relying on one search.
- Gathers sources: It searches, opens pages and follows useful leads. Perplexity says Research performs dozens of searches and reads hundreds of sources, but the number varies with the task, available sources and account.
- Analyzes and synthesizes: It compares findings and may use reasoning or coding capabilities where appropriate.
- Produces an output: It creates a report with citations and offers export or sharing options.
These are Perplexity’s product descriptions, not a guarantee of an exhaustive scholarly review. It may miss material, misunderstand a source or stop at information that is easy to discover.
Deep Research versus ordinary Perplexity search
| Ordinary search | Deep Research / Research mode |
|---|---|
| Best for a quick fact or a few sources | Best for complex, multi-part questions |
| Usually needs more user steering | Runs a multi-step investigation autonomously |
| Returns quickly | Usually takes several minutes |
| Compact answer with citations | Report-style synthesis with broader source gathering |
| You choose which pages to open | The system automatically searches and evaluates many pages |
| Efficient for one fact | Useful for comparisons, backgrounders and issue maps |
For a question such as “When was the regulation published?”, normal search is generally the efficient choice. Research mode earns its extra time when you need several sources reconciled or a structured first draft.
How to use it on the web, mobile or Mac
- Open Perplexity on the web, in its mobile app or in the Mac app.
- Open the mode selector beside the query field. Depending on the interface and rollout, it may be inside a + menu.
- Select Research or Deep Research, whichever label your account shows.
- Enter a specific brief with a date range, geography, source requirements and desired format.
- Let the searches and synthesis finish.
- Open the underlying citations and check important claims against the original documents.
- Export the result to a document or PDF, or share it as a Perplexity Page if that option appears.
If the mode is missing, sign in to the intended account, inspect the + menu, try the web app, and check plan or regional availability. Feature placement and access can vary by account and rollout. The current access guidance is in Perplexity’s Research mode Help Center article.
Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Prompts that produce useful reports
Specific constraints give the agent something testable to follow. A strong brief states the scope, evidence standard and output.
Comparison
“Compare the five-year total cost of ownership of three electric vehicles sold in the United States. Include purchase price, charging, insurance assumptions, maintenance and depreciation. Use manufacturer, government and insurer sources, state assumptions, and put disagreements in a separate table.”
Policy evidence
“Summarize evidence for and against this proposed policy from 2019–2026. Separate peer-reviewed studies, government data, stakeholder statements and commentary. Identify methodological limits and cite the original documents.”
Vendor selection
“Compare three project-management tools for a 25-person software team. Evaluate permissions, integrations, security documentation, data residency and current US pricing. Exclude affiliate roundups and flag any detail you cannot verify from a primary source.”
Do these 3 things before closing this tab:
1Fix the driver behind crashes, sound loss and screen glitches2Repair Windows errors before they cause bigger problems3Scan for outdated or missing drivers - takes under a minuteLiterature mapping
“Build a literature map of research on [topic] published from 2019 through 2026. Identify seminal papers, competing explanations, unresolved questions and studies that report negative or null results. Distinguish original research from reviews.”
Prompts such as “Tell me everything about AI” or “What is the best laptop?” are too broad to produce a dependable decision document. Medical, legal and financial questions also need explicit uncertainty instructions and independent review.
Free access, Pro access and model choice
Perplexity’s current Help Center says free users receive limited Research access and Pro subscribers receive extended access. It does not establish a universal quota in the documented material, so an exact daily or monthly number should not be assumed; limits can change by plan, account and date.
Users cannot manually select a model inside Research mode. Perplexity says the feature automatically chooses a combination of models. In its February 6, 2026 upgrade announcement, the company said Deep Research was running on Claude Opus 4.5 for Max users, with rollout to Pro users. That is a dated implementation detail, not a permanent specification for every account.
Perplexity also describes many tasks as completing in under three minutes, while the same Help Center material says generation can take around four to five minutes. Treat “a few minutes” as an approximate product claim, not a service-level promise.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What changed in February 2026
Perplexity’s February 6, 2026 changelog describes an upgraded Deep Research system using Perplexity’s search engine and sandbox infrastructure. It also describes model and plan-specific rollout details and introduced Model Council for higher-tier users. Model routing and availability can change, so these notes should be read as a snapshot of that release rather than an evergreen hardware specification.
Reliability: citations are evidence to audit, not proof
A cited report can still contain a mismatched citation, an outdated statistic, an overconfident inference, omitted counterevidence or a summary that changes the source’s meaning. Several pages repeating the same statement may simply reflect syndication rather than independent confirmation. Independent launch coverage has also raised concerns that research agents tend to favor readily discoverable material and can reduce a user’s engagement with the underlying evidence; that is a general risk, not proof that every Perplexity report fails in that way.
Use this verification checklist
- Open the original source instead of relying on the citation preview.
- Check that it supports the exact sentence, not merely the same topic.
- Verify publication date, jurisdiction, population and geographic scope.
- Prefer primary documents for statistics, laws, prices, product specifications, filings and medical claims.
- Ask for counterarguments, failed studies, recalls, lawsuits, regulatory warnings and other negative evidence where relevant.
- Request a table separating verified facts, inferences and unresolved uncertainty.
- Re-run consequential claims as narrow, targeted searches.
Do not use an unverified report as the sole basis for a medical, legal, investment, safety or regulatory decision. A research agent gathers and organizes evidence; it does not assume responsibility for your judgment.
When Research mode is a good fit
- Current web information is central to the assignment.
- You need a comparison or synthesis across many sources.
- You want citations embedded in a shareable first-pass report.
- You are mapping an issue before doing deeper primary-source work.
- You are researching finance, marketing, technology, current affairs, health, biography or travel and can verify consequential details.
When another workflow is better
- The answer is one easily checked fact; ordinary search is faster.
- You need a reproducible systematic-review method or guaranteed completeness.
- The key evidence is paywalled, proprietary, local or poorly indexed.
- You require exact quotations or controlling legal language.
- Your work depends on a private document corpus and strict source controls.
When comparing competing research agents, examine source restrictions, claim-level citation quality, freshness, uploaded-document support, visibility into the search plan, steerability, export formats, quotas, privacy terms and reproducibility. No benchmark or model label establishes universal superiority across every subject.
Consumer plans and the API are different choices
Perplexity Pro and Max are consumer subscriptions; the live plan page at perplexity.ai is the right place to check current prices and allowances. The API is a separate developer product. Perplexity’s API changelog documents an asynchronous Sonar Deep Research API for longer research-intensive jobs, with submission and later retrieval. It suits teams building repeatable pipelines, not someone who occasionally wants a browser report.
Verdict
Perplexity Deep Research—now commonly labeled Research mode—is worth trying for broad, current, multi-source web research, comparisons and issue mapping. The free tier may cover occasional users; frequent researchers should compare live Pro or Max limits with alternatives such as ChatGPT’s Deep Research. In every tier, treat the generated report as a well-organized starting point and verify the sources before relying on its conclusions.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
