Provider pricing

Turbopuffer AI API pricing

Turbopuffer is a serverless search engine that keeps namespaces on object storage and only loads hot data into memory. The architecture delivers 10-100x lower cost than traditional vector databases for read-heavy workloads.

Founded2023
HeadquartersSan Francisco, CA
Tracked modelsDirectory

Best for

Teams with many tenants or cold vector data who need cheap, scalable search.

Strengths

  • Object-storage-native — orders-of-magnitude cheaper than in-memory DBs
  • Vector + full-text + filter in one query
  • Per-namespace isolation for multi-tenant apps
  • Usage-based pricing with no idle cost