turbopuffer is a serverless vector and full-text search engine built from the ground up on object storage, designed for speed, high scalability, and cost efficiency.

It supports hybrid search combining vector and BM25 methods, handles massive workloads with production deployments exceeding 150 billion documents and thousands of queries per second, and offers pricing that is significantly lower than traditional solutions. The system is usage-based, cloud-native, and targets AI-driven and large-scale data retrieval use cases.