← Back to glossary
+Suggest a term
Tool·Builder Tools·Added 1 day ago

Perplexity

Also known as: Perplexity AI, Perplexity search

An AI-powered answer engine that searches the web in real time and returns direct, cited answers rather than a list of links. Founded in 2022 and valued at approximately $20 billion as of September 2025, Perplexity processes over 1 billion queries per month and has expanded into agentic computer use with its Computer product.

Perplexity AI was founded in August 2022 by Aravind Srinivas (formerly at OpenAI and Google DeepMind), Denis Yarats, Johnny Ho, and Andy Konwinski. The product launched as a search engine that uses large language models to synthesize results from live web searches and return a single cited answer, rather than a page of links for the user to parse. It positioned itself as an 'answer engine' and attracted early adopters who found it faster and more direct than traditional search for research and technical questions.

The product has expanded significantly. As of May 2026, Perplexity has over 100 million monthly active users across all products and $450+ million in annualized recurring revenue. Its product surface includes the core search/answer engine, Deep Research (a premium feature that conducts multi-step research across hundreds of sources and produces a structured report), the Comet browser (an AI-native browser built on Chromium), and Perplexity Computer (an agentic AI launched February 2026 that executes complex multi-step tasks autonomously). The company transitioned to a subscription-first model in February 2026, dropping its AI-integrated advertising strategy.

For builders, Perplexity matters in two ways: as a product example of how to do AI-powered search and grounded answers well (with citations and real-time data as first-class features), and through its developer API (the Sonar API) which gives programmatic access to Perplexity's search and answer capabilities. Perplexity also released open-source embedding models (pplx-embed-v1) in February 2026, positioned as state-of-the-art for dense retrieval tasks.

This definition is AI-generated and refreshed weekly. It may contain inaccuracies. Use your own judgment, especially for production decisions.
Related terms
RAGGroundingDeep ResearchEmbeddings