What Is Perplexity?
Perplexity is an AI “answer engine,” a cross between a search engine and a chatbot. You ask a question in plain language, and instead of a list of blue links or an unsourced paragraph, you get a direct, written answer with citations linking to the sources it used.
That citation-first approach is the whole point. Where a standard chatbot can state something confidently and wrongly, Perplexity shows its working, letting you click through and verify. It combines live web search with large language models, and on the Pro plan, you can choose between models like GPT and Claude.
It suits researchers, students, professionals, and anyone who wants current, checkable answers rather than either a pile of links or an AI’s unsourced say-so.
How Does Perplexity Work?
You type a question. Perplexity searches the web in real time, reads the relevant pages, and synthesises a concise answer with numbered citations. You can ask follow-ups conversationally, and it keeps the thread’s context.
Beyond basic search, it offers Focus modes (academic, writing, and more), file uploads for asking questions about your own documents, and Labs for deeper multi-step research tasks. Its agentic browser, Comet, goes further; it can take actions on your behalf, including shopping, which is where the controversy begins.
The Amazon Lawsuit and Trust Questions
If you’ve searched “Perplexity controversy,” this is what you found. In November 2025, Amazon sued Perplexity over its Comet browser’s shopping agent. Amazon alleges Comet accessed customer Amazon accounts without authorization and disguised its automated activity as a human using a normal browser, which it argues breaches the Computer Fraud and Abuse Act and creates security risks.
Perplexity pushed back hard, calling the suit a “bully tactic” and arguing that users should be free to use any AI assistant they choose, and that AI agents acting on a user’s behalf should have the same access a person does.
In March 2026, a judge granted Amazon a preliminary injunction blocking Comet from password-protected areas of Amazon. Perplexity appealed, and an appeals court granted a temporary pause. As of 2026, the case is ongoing with no final ruling, and its outcome could shape the rules for AI agents across the whole web.
Key Perplexity Features
- Cited Answers
- Every response includes inline citations linking directly to the sources it drew from, so users can trace claims back to original articles, papers, or websites instead of taking the answer at face value. This makes it easier to verify accuracy, spot bias, and dig deeper into a topic without running a separate search.
- Real-Time Web Search
- Unlike models limited to a training cutoff, Perplexity pulls live results from the web for each query, so answers reflect current events, recent data, and up-to-date information. This makes it especially useful for news, pricing, product comparisons, and anything time-sensitive.
- Model Choice (Pro)
- Pro users can switch between leading underlying models, including options from OpenAI and Anthropic, depending on the task at hand. This lets users pick a model suited to reasoning, creative writing, or factual lookup rather than being locked into a single provider’s strengths and limitations.
- Focus Modes
- Users can narrow searches to specific source types, such as academic papers, Reddit discussions, YouTube, or general web results. This filters out noise and surfaces more relevant, higher-quality sources depending on whether the task is research, casual browsing, or technical troubleshooting.
- File Uploads
- Users can upload their own PDFs, documents, or images and ask direct questions about the content. Perplexity reads and reasons over the uploaded material, making it useful for summarising reports, extracting data, or getting explanations of dense or technical documents.
- Labs and Comet
- Labs handles deeper, multi-step research tasks that go beyond a single search-and-answer exchange, while Comet is an agentic browser that can navigate websites and complete actions on a user’s behalf. Together, they extend Perplexity from a search tool into something closer to a research and task-automation assistant.
Imperial AI Tools Feedback
Perplexity does one thing better than almost anyone: it answers questions with sources you can check. In an era of confident AI hallucination, that citation-first design is genuinely valuable, and the real-time search keeps it current where static chatbots fall behind. For research, fact-finding, and staying informed, it’s excellent, and the free tier is strong enough to prove it.
The honest caveats are twofold. First, it’s still AI; it can misread or misattribute a source, so the citations are there to be checked, not trusted blindly. Second, the Amazon lawsuit over its Comet browser raises fair questions about how aggressively the company pursues agentic access. That dispute doesn’t affect core search, but it’s a reasonable thing to factor into how much you trust the broader platform.
Our view: one of the best AI research tools available, and for sourced answers it often beats a general chatbot. Use it for what it’s best at: checkable research, click through its citations, and approach the agentic Comet features with more caution until the legal picture clears.
Suggestions For Improvement
- Continue improving citation accuracy to reduce misattribution.
- Be transparent about the Comet agent’s access and data handling.
- Clarify how user data is used across free and Pro tiers.
- Maintain fair relationships with the publishers it cites.
- Keep answer quality consistent as it scales and adds features.
