Perplexity Partners With Vespa.ai to Bring its Search Function In-House

Perplexity, the AI-powered answer engine, has partnered with Vespa.ai to bring its search infrastructure in-house. This strategic shift enhances the speed, accuracy, and scalability of Perplexity's search capabilities, leveraging Vespa's robust platform for building large-scale, real-time AI applications powered by big data.

“The recipe: 1. Solve Search. 2. Use it to solve everything else,” said Aravind Srinivas, CEO of Perplexity, emphasizing the centrality of search in the AI space.

Perplexity’s approach to delivering direct, sourced answers is built on a scalable Retrieval-Augmented Generation (RAG) architecture. This framework efficiently accesses data from the web, internal databases, and user-uploaded files. Powered by Vespa, Perplexity now serves accurate, near-instant answers to more than 15 million monthly users, managing over 100 million weekly queries.

By leveraging Vespa.ai’s managed platform, Perplexity ensures low-latency and real-time information retrieval from extensive datasets. Vespa.ai provides the flexibility, speed, and reliability essential for high-performance conversational AI experiences.

“The project with Perplexity has been an exciting opportunity to showcase what Vespa’s platform can do,” said Jon Bratseth, CEO of Vespa.ai. “We’re thrilled to support Perplexity’s mission by powering their next-generation search capabilities.”

Key Features of the Vespa-Powered Search Platform:

  • High-Performance Vector and Text Search: Vespa’s advanced vector search combined with powerful text search enables fast, accurate information retrieval from massive datasets.
  • Scalability and Reliability: Vespa’s distributed architecture ensures seamless scalability and high availability as user demand continues to grow.
  • Machine-Learned Ranking: Native distributed ML ranking inference allows for sophisticated, signal-based relevance tuning across billions of documents.
  • Cost Efficiency: Vespa’s resource-efficient infrastructure helps Perplexity maintain performance while optimizing operational costs.

This collaboration underscores the transformative role of RAG systems in modern AI applications and highlights the value of high-performance hybrid search engines. With Vespa.ai, Perplexity is well-positioned to redefine how users interact with and access information at scale.