Turbopuffer

About

Turbopuffer develops a vector and full-text search database built directly on object storage. Their core technology delivers scalable and low-latency similarity search, claiming a 10x cost reduction compared to conventional vector databases. Currently serving production systems handling over 4 terabytes of data and 10 million writes per second, Turbopuffer targets companies building AI applications like semantic search and recommendation systems requiring high-performance search at scale.

Focus Areas

Technology Focus