Perplexity AI

Perplexity AI is a privately held American artificial-intelligence software company best known for a conversational search service that combines large language models with real-time web retrieval. Instead of presenting only a ranked list of links, the service attempts to interpret a question, gather relevant online material, and produce a direct answer with citations to the sources it used. This positioning places Perplexity between a conventional search engine, an answer engine, and a generative-AI assistant. The company was founded in August 2022 by Aravind Srinivas, Denis Yarats, Johnny Ho, and Andy Konwinski. Its principal search product launched publicly on December 7, 2022. The founders brought experience in machine learning, artificial intelligence, software engineering, and backend systems. Perplexity initially attracted attention by emphasizing current information, source links, conversational follow-up questions, and a relatively simple interface compared with general-purpose chatbot products. Perplexity operates a freemium model. Its public search experience is available without charge, while Perplexity Pro provides expanded access to advanced models, higher usage limits, document analysis, and additional research capabilities. The platform can route requests to different underlying models, including models developed by Perplexity and models supplied by other AI companies. Its Sonar family is positioned as a search-oriented model and is used in the company’s developer-facing search infrastructure. The company has also introduced R1 1776, an internally developed model based on DeepSeek R1. The brand has expanded beyond a web search page. Products and features associated with Perplexity include a browser extension, iOS and Android applications, Perplexity Pages for creating structured content, finance tools, shopping recommendations, an AI Assistant for carrying out actions across applications, Internal Knowledge Search for querying company documents alongside the web, and a Search API for developers. In 2025 it introduced Comet, a Chromium-based browser integrating Perplexity’s search and task capabilities. The company has also pursued enterprise use cases, partnerships with publishers, and integrations with external platforms. Perplexity’s growth has been accompanied by significant controversy. Publishers and media companies, including Forbes, Dow Jones, the New York Post, The New York Times, the BBC, and Japanese newspaper groups, have alleged that the service reproduces or accesses copyrighted material without adequate permission, licensing, attribution, or compensation. Some allegations also concern fabricated or incorrectly attributed quotations in generated responses. Investigations by Wired, independent researchers, and Cloudflare further alleged that Perplexity-associated crawlers used undisclosed infrastructure or spoofed user-agent information to access sites that attempted to restrict automated collection. Perplexity has disputed aspects of those claims and has argued that its system aggregates information while expressing willingness to develop revenue-sharing arrangements with publishers. The company remains private and has been financed by venture investors and strategic backers, including technology and AI-focused investors. Its principal public face is co-founder and chief executive Aravind Srinivas. Perplexity’s long-term position depends on whether it can maintain answer quality and user trust while resolving publisher relationships, copyright questions, crawler-governance concerns, infrastructure costs, and competition from established search companies and general-purpose AI platforms.

Perplexity AI was established in August 2022 by Aravind Srinivas, Denis Yarats, Johnny Ho, and Andy Konwinski. The founding team combined experience in machine learning, artificial intelligence research, backend engineering, and software systems. Their objective was to create a search experience in which a user could ask a question in natural language and receive a concise, conversational response supported by links to source material. The company released its principal search engine on December 7, 2022. Its early differentiation came from combining language-model generation with live web retrieval. Rather than treating the language model as an isolated knowledge base, Perplexity designed the service to consult current online content and expose citations alongside its answers. This approach made the product useful for research, news-oriented questions, explanations, and follow-up conversations, although it also created risks involving inaccurate synthesis, incomplete citations, and the reproduction of publisher content. Perplexity expanded access through web, mobile, and browser products. It released applications for iOS and Android and developed a Chrome extension. The company introduced a paid Pro tier with access to stronger models, increased capabilities, document uploads, and additional research functions. It also developed enterprise-oriented features that allow users to search internal files together with public web material. Supported files include common office and document formats, enabling the product to serve as a workplace knowledge-search tool rather than solely a public web interface. The company broadened its commercial scope through developer infrastructure and vertical features. Its Search API provides programmatic access to Perplexity’s retrieval capabilities, while the Sonar model family is designed for search-related workloads. Finance features support market and company research, and Shopping Hub applies conversational recommendation to product discovery and purchasing. In January 2025, Perplexity introduced an Assistant designed to carry out actions across applications and to use multimodal inputs such as a phone camera. In July 2025, it introduced Comet, a Chromium-based browser that placed Perplexity’s search and task tools within the browsing environment. Perplexity’s growth brought substantial attention from investors, publishers, technology companies, and competitors in search and generative AI. The company remained privately held and raised capital from a range of venture and technology investors. Its chief executive, Aravind Srinivas, became the principal public spokesperson for the business and regularly described the company as an alternative to conventional search and advertising-led information discovery. The central challenge in Perplexity’s history has been its relationship with the open web. In 2024, Forbes criticized the company over an answer that allegedly reproduced material from a proprietary article. Dow Jones and the New York Post sued over alleged copyright infringement and inaccurate attribution, while The New York Times sent a cease-and-desist notice. Perplexity responded that its system aggregates information rather than plagiarizing it and stated that it was open to commercial arrangements with publishers. Further disputes followed in 2025. The BBC demanded that Perplexity stop using its material without authorization and address alleged retention of BBC content. Japanese newspaper companies filed lawsuits alleging extensive unauthorized use of their articles, and Reddit later brought a claim concerning alleged data scraping. Separately, investigations by Wired, independent researchers, and Cloudflare questioned whether Perplexity-associated crawlers consistently respected robots.txt and published crawler-identification practices. Perplexity rejected important parts of those allegations. These disputes have made licensing, attribution, crawler transparency, and the legal status of AI search central to the brand’s public identity. Perplexity is therefore best understood as both an AI application company and a search-infrastructure company. Its brand promise rests on speed, direct answers, current information, and visible sourcing. Its future development depends on improving factual reliability, expanding paid and enterprise use, controlling infrastructure costs, and establishing sustainable relationships with the publishers and websites whose content supplies the information layer behind its answers.

• 2025

  Assistant launched

  [1]Perplexity AIen.wikipedia.org

  Perplexity released a multimodal assistant intended to perform actions across applications and maintain context across tasks.

• 2025

  Comet browser launched

  [1]Perplexity AIen.wikipedia.org

  The company launched Comet, a Chromium-based browser integrating Perplexity search and AI task assistance.

• 2024

  Publisher program announced

  [1]Perplexity AIen.wikipedia.org

  Perplexity announced a publisher program intended to share advertising revenue with participating content partners.

• 2024

  Finance tools introduced

  [1]Perplexity AIen.wikipedia.org

  The service added features for market-price tracking, earnings information, peer comparisons, and basic financial research.

• 2024

  Shopping Hub launched

  [1]Perplexity AIen.wikipedia.org

  Perplexity introduced an AI-powered shopping experience for recommendations and product discovery.

• 2023

  Early public adoption reported

  [1]Perplexity AIen.wikipedia.org

  Perplexity reported substantial early visitor activity as the answer-engine category began attracting wider public attention.

• 2022

  Company founded

  [1]Perplexity AIen.wikipedia.org

  Perplexity AI was founded in August by Aravind Srinivas, Denis Yarats, Johnny Ho, and Andy Konwinski.

• 2022

  Main search engine launched

  [1]Perplexity AIen.wikipedia.org

  The company launched its conversational search engine on December 7, combining generative answers with web citations.

A premium, citation-oriented AI answer engine that combines conversational interaction, large language models, and real-time web search.

• Conversational AI search engine · 2022

  Perplexity Search

  Perplexity’s core public service accepts natural-language questions and generates synthesized responses by combining large language models with real-time web retrieval. Answers normally include citations or links to the material consulted, allowing users to inspect supporting sources. The product follows a freemium model and is available through the web, mobile applications, and browser integrations. Its principal differentiation is the attempt to provide a researched answer rather than merely a page of search results.

• Paid AI search subscription

  Perplexity Pro

  Perplexity Pro is the company’s paid tier for users who need greater access to advanced models and research functions. Depending on the product configuration, it supports model selection, higher usage allowances, document uploads, and access to API or enterprise-oriented capabilities. The service is positioned for intensive individual research and professional use, extending the free answer-engine experience with broader tools and controls.

• Search-oriented AI model and API

  Sonar

  Sonar is Perplexity’s search-oriented model family and a foundation of its developer-facing retrieval infrastructure. It is designed to support applications that need current information, generated responses, and web-grounded results. Perplexity has described Sonar as being based on Meta’s Llama model family. The associated Search API gives developers programmatic access to Perplexity’s search capabilities.

• Enterprise knowledge management

  Internal Knowledge Search

  Internal Knowledge Search allows Pro and Enterprise Pro users to search public web content and uploaded internal material in a combined workflow. It supports common business file types such as PDF, Word, and Excel documents. The feature is intended to help employees answer questions using both external research and organizational records, making Perplexity relevant to enterprise knowledge retrieval and document analysis.

• AI shopping and product discovery · 2024

  Shopping Hub

  Shopping Hub applies Perplexity’s conversational interface to product research. It generates recommendations and helps users compare or discover products, with the broader objective of supporting purchases inside the search experience. The product extends the brand from information retrieval into commerce while retaining its emphasis on natural-language questions and synthesized guidance.

• Multimodal AI assistant · 2025

  Perplexity Assistant

  The Perplexity Assistant is designed to move from answering questions to carrying out actions. It can maintain context, interact with selected applications, and support tasks such as arranging a ride or identifying music. Its multimodal functions can use a phone camera or on-screen content as input. Perplexity has acknowledged that some functions remain experimental and may require user-configured workarounds.

• AI web browser · 2025

  Comet

  Comet is a Chromium-based browser built around integrated Perplexity search. It can assist with article summaries, image descriptions, topic research, and drafting communications while the user browses. Its strategic purpose is to place AI search directly inside the browsing environment rather than requiring users to switch between a search page and other applications.

• Large language model

  R1 1776

  R1 1776 is a Perplexity-developed model based on DeepSeek R1. It represents the company’s effort to offer proprietary or customized model capabilities alongside access to external model providers. The model is associated with Perplexity’s broader strategy of controlling parts of the technology stack behind its search and answer products.

• 2025

  BBC and Japanese publisher legal disputes

  The BBC demanded that Perplexity stop unauthorized use of its material and address alleged retention of BBC content. Yomiuri Shimbun, The Asahi Shimbun, and Nikkei separately brought claims alleging unauthorized use of large quantities of newspaper content.

  Source

• 2025

  Allegations concerning stealth web crawlers

  Investigations by Wired and independent researchers, followed by Cloudflare research, alleged that Perplexity-associated crawlers used undisclosed IP infrastructure or spoofed user-agent strings and did not consistently honor robots.txt restrictions. Perplexity denied key elements of the allegations and criticized Cloudflare’s presentation.

  Source

• 2025

  Trademark lawsuit by Perplexity Solved Solutions

  Perplexity Solved Solutions sued Perplexity AI, alleging that the latter’s use of the Perplexity name infringed a registered trademark and could confuse consumers. The claimant reportedly said that an earlier offer to purchase the mark had not been accepted.

  Source

• 2025

  Reddit data-scraping lawsuit

  Reddit sued Perplexity and other companies, alleging unlawful scraping and use of Reddit data. The dispute added a platform-data dimension to the company’s existing publisher and crawler controversies.

  Source

• 2024

  Copyright and attribution disputes with publishers

  Forbes, Dow Jones, the New York Post, and The New York Times raised allegations that Perplexity reproduced, accessed, or relied on publisher content without adequate authorization or attribution. The Dow Jones and New York Post litigation also alleged that generated answers included inaccurate quotations attributed to news coverage. Perplexity disputed the characterization of its system as plagiarism and said it was open to revenue-sharing relationships.

  Source

Read more

• 2024Publisher revenue-sharing program

  United States · Global

  Perplexity announced a program intended to share advertising revenue with publishers whose content contributed to the company’s answer ecosystem. The initiative was presented as a way to create a commercial relationship with content producers amid growing concerns about AI search and publisher economics.

  Outcome. The program did not eliminate broader disputes over scraping, attribution, licensing, and copyright.

  Source

• 2025

  Extend the brand into browser-based AI assistanceProduct launch

  Background

  Perplexity sought to make its search and task functions available throughout the browsing experience.

  Action taken

  The company launched Comet, a Chromium-based browser incorporating search, research, summarization, and task assistance.

  Aftermath

  The launch broadened Perplexity’s competitive scope from search software toward browser and agent experiences.

  Source

• 2024

  Pursue publisher partnerships and revenue sharingStrategy

  Background

  Publishers questioned whether AI-generated answers could use their reporting without creating licensing or compensation obligations.

  Action taken

  Perplexity announced a publisher program intended to share advertising revenue with participating partners.

  Aftermath

  The initiative coexisted with continuing copyright, attribution, and web-crawling disputes.

  Source

• 2022

  Launch a citation-oriented conversational search engineProduct launch

  Background

  The founders sought to make web research more direct and conversational than a conventional list of ranked links.

  Action taken

  Perplexity launched a service that uses language models to synthesize web-retrieved information and display supporting citations.

  Aftermath

  The model became the foundation for the company’s consumer, enterprise, and developer products.

  Source

• 2026.10.20

  Reported2026.10.20

  Valarian and Perplexity announce sovereign AI partnership[3]Valarian与Perplexity宣布主权人工智能合作finanznachrichten.de

  Valarian announced a partnership with Perplexity to make Perplexity’s AI tool suite available within the compute boundaries of organizations using Valarian’s ACRA environment.

• Unconfirmed

  Reported2026.10.08

  HP to add Perplexity Computer to consumer PCs with one-month Perplexity Pro trial[4]惠普消费级PC将加入Perplexity Computer并提供一个月Perplexity Pro试用hothardware.com

  HP is adding Perplexity Computer to its consumer PCs, with eligible systems receiving a one-month Perplexity Pro trial beginning in November. HP says the software is intended to help users research information, complete tasks, and create content using the local compute in its new RTX Spark systems; Microsoft's onboarding experience for RTX Spark machines also includes Perplexity as an agent that can be set up.

• 2026.08.25

  Reported2026.08.26

  Perplexity AI launches Portable Computer on-device AI agent[5]Perplexity AI 发布 Portable Computer 端侧 AI 智能体bnext.com.tw

  Perplexity AI Inc. introduced Portable Computer, an on-device AI agent that runs on desktops equipped with Nvidia silicon, including the Nvidia DGX Spark device. The tool is an on-device version of the cloud-based Perplexity Computer agent and uses the open-source Qwen 3.8 27B language model along with a Perplexity-optimized PPLX 27B version.

• Unconfirmed

  Reported2026.08.25

  Perplexity introduces Portable Computer local AI agent[6]Perplexity 推出 Portable Computer 本地 AI 代理digg.com

  Perplexity posted that Portable Computer is a local-first agent for private and cost-effective work, featuring an on-device 27B model whose harness scores 82.6% on real knowledge work, beating open-source harnesses Pi and Hermes, according to the company. Perplexity said its post-trained PPLX 27B reaches 85.4%, and Aravind Srinivas noted advisor escalation to a cloud frontier model lifts Terminal Bench 2.1 scores from 59.6% to 73.0% at $0.415 per rollout.

Read more

1. 1Perplexity AIen.wikipedia.org
2. 2Perplexity AI official hubperplexity.ai
3. 3Valarian与Perplexity宣布主权人工智能合作finanznachrichten.de
4. 4惠普消费级PC将加入Perplexity Computer并提供一个月Perplexity Pro试用hothardware.com
5. 5Perplexity AI 发布 Portable Computer 端侧 AI 智能体bnext.com.tw
6. 6Perplexity 推出 Portable Computer 本地 AI 代理digg.com
7. 7Perplexity 攜手輝達推出 Portable Computer，AI 代理全程在地端執行cdn.technews.tw
8. 8Perplexity AI launches Portable Computer on-device AI agent - SiliconANGLEsiliconangle.com
9. 9Perplexity Launches Local AI Model That Will Run on Your GPU Instead of the Cloudgizmodo.com