SAN FRANCISCO, April 9 — Perplexity AI on Thursday introduced Pro Search, a higher-tier offering layered on top of its existing $20-per-month consumer subscription, that lets users restrict queries to a curated allowlist of publishers and academic sources. The feature, the company said, is a response to enterprise customers asking for tighter provenance controls on AI-generated answers.
The launch is Perplexity’s most explicit move yet toward the enterprise market, where competitor offerings from OpenAI, Anthropic, and the incumbent search vendors have intensified pressure on the four-year-old startup’s consumer-search positioning. Pro Search is priced at $40 per month for individuals and from $50 per seat per month for teams of ten or more, with a higher-tier “Verified Sources” enterprise plan available on request.
“What we heard from customers, particularly in regulated industries, is that the answer is only useful if you can constrain the source set the answer is drawn from,” said Dmitri Srinivas, Perplexity’s vice president of product, in an interview with Consumer Tech Wire. “Pro Search lets a healthcare team restrict queries to peer-reviewed journals and FDA filings, or a finance team restrict queries to SEC filings and a defined list of publications, and then returns answers with citations to that constrained set only.”
Model picker, agentic mode
Pro Search also expands the model picker that has become a Perplexity hallmark. Subscribers can now route queries through Claude Opus 4, GPT-5, Gemini 3 Pro, or Perplexity’s own Sonar Large model, which the company said has been retrained on a refreshed web index. An agentic-research mode, branded as “Pro Deep Research,” runs multi-step queries that the company said can produce “report-length” outputs of up to 8,000 words with inline citations.
The Deep Research feature is rate-limited to ten runs per day on the $40 tier and unlimited on the enterprise plan.
Competition and the publisher question
The launch comes as the relationship between AI-search companies and news publishers has continued to deteriorate. The New York Times, Condé Nast, and a coalition of regional news organizations have sued OpenAI and Perplexity in separate actions over training-data and answer-attribution practices, with several suits still working through federal court.
Perplexity said Pro Search will only return answers from publishers participating in its publisher revenue-share program, which the company said now numbers more than 80 outlets, including Time, Fortune, Der Spiegel, and The Texas Tribune.
“The publisher question is structurally unresolved across the AI-search category,” said Lukas Brennan-Kowalski, an analyst at the media-research firm Toolkit. “What Perplexity is doing with Pro Search is the most coherent answer anyone has put on the table — pay publishers, restrict queries to participating publishers, surface attribution. Whether the economics work at $40 a month is a separate question.”
A Perplexity spokesperson said the company will publish quarterly reports on publisher revenue-share payouts beginning in July.
Asari Whitfield-Asari covers AI tools and developer infrastructure for Consumer Tech Wire.