Case Study: Perplexity achieves faster, more accurate AI search with Vespa

Perplexity powers 780M+ monthly queries with Vespa

Perplexity, an AI-powered answer engine, faced the challenge of building a retrieval system capable of finding the most relevant information to provide accurate, cited answers. Their goal required a platform that could combine keyword and semantic search with sophisticated filtering and ranking at a massive consumer scale, as existing tools like vector databases and traditional search backends were insufficient. They chose the Vespa platform to solve this.

Vespa provided a single, integrated solution that combined comprehensive hybrid retrieval, real-time continuous indexing, and multi-phase machine learning ranking. This allowed Perplexity to build a high-performance search infrastructure that processes over 780 million queries per month with 20% monthly growth, ensuring their AI receives the most relevant, fresh context to generate trustworthy answers for its users.

Perplexity