Case Study: Plaud achieves billion-scale agent memory and sub-200 ms retrieval with Zilliz Cloud

Plaud powers billion-scale agent memory with Zilliz Cloud and sub-200 ms recall

Plaud, the world's top AI note-taking brand, faced the challenge of building a consumer-scale AI memory system. Their existing OpenSearch infrastructure could not handle the explosive growth of data or the demanding requirements of billion-scale semantic and hybrid search needed for their agentic RAG workload. They needed a solution that could manage massive, continuous data growth, ensure user privacy, and provide always-on global access without downtime.

By implementing Zilliz Cloud, Plaud gained a scalable vector database foundation. The solution enabled them to power their ContextOS retrieval layer, handling billions of vectors with average recall latencies under 200ms. This resulted in the ability to run billion-scale historical data backfills in days with zero service interruption, all while maintaining strict data residency and privacy controls. Zilliz Cloud provided a managed service that eliminated operational overhead, allowing Plaud's team to focus on product innovation instead of infrastructure.

Plaud