Perplexity AI
Perplexity AI是一家专注于人工智能搜索领域的科技公司，总部位于美国旧金山。其核心产品为基于大语言模型的对话式搜索引擎，能够对用户查询进行实时网络检索并以自然语言生成综合性答案，同时提供信息来源引用。公司致力于以AI技术重构传统搜索体验，产品覆盖网页端、移动端及浏览器等多个平台。
Core Facts
Perplexity AI was founded in 2022, co-founded by former OpenAI researcher Aravind Srinivas, and is positioned as an "answer engine."
Perplexity AI's underlying architecture uses RAG (Retrieval-Augmented Generation) technology, crawling web content in real time and combining it with LLM output to produce structured answers with cited sources.
Perplexity was founded in 2022 by Aravind Srinivas, a former OpenAI researcher.
Timeline (last 90 days)
Perplexity AI released the pplx-embed-v2-late series of models, comprising two parameter versions, 0.6B and 9B.
Both models have been made public on the Perplexity AI organization page on Hugging Face.
The author used Perplexity AI to research the pricing of similar tools, and found that competitors' monthly fees generally hover around 20 USD.
Perplexity AI has launched a shopping mode with built-in purchase buttons.
Perplexity emphasizes that the URLs in its index are all high-quality content accompanied by high-quality summary information.
Hermes is Perplexity's internal platform.
Perplexity has set a target of reaching a search index of 1 trillion URLs by the end of the year.
Perplexity's current search index already covers more than 450 billion URLs.
Perplexity AI announced the integration of search functionality into the Hermes platform.
Performing content preprocessing and structuring on a search index can reduce the computational burden of real-time inference and improve response speed.
14 more timeline events
All Facts (20)
Perplexity AI was founded in 2022, co-founded by former OpenAI researcher Aravind Srinivas, and is positioned as an "answer engine."
90%VerifiedPerplexity AI's underlying architecture uses RAG (Retrieval-Augmented Generation) technology, crawling web content in real time and combining it with LLM output to produce structured answers with cited sources.
90%VerifiedPerplexity was founded in 2022 by Aravind Srinivas, a former OpenAI researcher.
85%VerifiedPerplexity AI rose to prominence in late 2022, directly challenging traditional search with its positioning as an 'answer engine'.
85%VerifiedPerplexity AI was founded in 2022.
75%VerifiedPerplexity AI's monthly active users surpassed 15 million in 2024.
75%VerifiedThe "answer engine" paradigm represented by Perplexity abandons returning lists of web pages and instead directly generates natural language answers with cited sources, which is essentially the productization of RAG technology.
65%VerifiedIn 2024, The New York Times and several other media outlets accused Perplexity of unauthorized mass crawling of paid content to generate summaries, sparking copyright disputes
65%UnverifiedPerplexity AI positions itself as an answer engine rather than a search engine, directly synthesizing answers from multiple sources and providing citation links.
60%UnverifiedPerplexity AI has launched a shopping mode with built-in purchase buttons.
50%UnverifiedHermes is Perplexity's internal platform.
50%UnverifiedPerforming content preprocessing and structuring on a search index can reduce the computational burden of real-time inference and improve response speed.
50%UnverifiedMaintaining a large-scale index requires enormous computational and storage resources.
50%UnverifiedPerplexity emphasizes that the URLs in its index are all high-quality content accompanied by high-quality summary information.
50%UnverifiedPortable Computer is a fully local runtime version of Perplexity Computer, allowing users to run Perplexity's AI capabilities on their own devices without relying on cloud servers.
50%UnverifiedPerplexity AI adopts a real-time search plus annotated citation model, listing the source of each argument with a rigor close to academic citation standards.
50%UnverifiedChatGPT's web browsing feature annotates specific source URLs, Perplexity AI adopts a real-time search plus annotated citation model, and Claude in certain scenarios tends to synthesize information from multiple sources without explicit annotation.
50%UnverifiedMerging multiple small files can spread the fixed costs of initialization, format parsing, and index construction across each upload, maximizing value within limited quotas.
50%UnverifiedPerplexity is positioned as an Answer Engine that directly generates comprehensive answers with cited sources, rather than returning a list of links.
50%UnverifiedBoth models have been made public on the Perplexity AI organization page on Hugging Face.
50%