Perplexity Patents beta version launched, featuring AI agents, natural language patent search, and breaking down barriers to traditional keyword search.
Perplexity recently released a new AI patent research tool called "Perplexity Patents," which is currently in beta testing and available for free trial to all users. Positioned as an "AI patent research agent," this tool aims to allow the general public to quickly find complex patent information using natural language commands, significantly lowering the barrier to entry for traditional patent searches. Addressing the pain points of high barriers in traditional patent searches, Perplexity points out that for decades, professional patent searches have heavily relied on systems requiring extensive training and familiarity with precise keyword combinations and complex syntax. General public search tools have limited functionality, while professional platforms are expensive; these limitations have resulted in patent information being almost exclusively in the hands of experts. Adopting a "Citation-First" Model and Supporting Conceptual Searches The core feature of Perplexity Patents is its "Citation-first" operating model. Unlike most AIs that directly generate answers based on model predictions, Perplexity Patents first finds the sources of citations (patent documents) before providing answers, thereby ensuring the transparency and credibility of the results. Users simply need to ask questions in natural language (e.g., "Are there any patents for language learning AI?", or "Since 2024...").



