Perplexity is an answer engine made by Perplexity AI: an AI search service that answers a question in conversational text instead of a list of links. It combines real-time web search with large language models (neural networks trained on large amounts of text, which generate text). These include its in-house model Sonar and models from OpenAI, Anthropic and Google.
For each question, Perplexity interprets the request, searches the web at that moment, and summarizes what the sources say. Each answer has numbered citations that link to the original sources, so a reader can check them. Pro Search handles multi-step questions in more depth, and Research mode runs dozens of searches, reads hundreds of sources and writes a report in two to four minutes. Through the Agent API, a program gets such a web-grounded answer, with citations, in one call.
The service has a free Standard plan with limits and paid subscriptions: Pro, Max, and Enterprise plans priced per seat. The API is billed per use, per token (a piece of a word) of input and output, plus a fee for each web search call.