Turbopuffer
Serverless vector and full-text search engine built on object storage
About Turbopuffer
Turbopuffer provides a serverless search database that stores data on object storage with caching to deliver vector and full-text search at lower cost. It is designed for large-scale retrieval workloads such as AI search and RAG. It is offered as a managed multi-tenant or dedicated service.
Common use cases
- RAG applications
- Large-scale semantic search
- Code and document search
- AI product search
Products and services
turbopuffer
Serverless vector and full-text search database.
Key features
- Object-storage-based
- Vector search
- Full-text search
- Serverless scaling
- Namespace isolation
- Low-cost storage
Integrations
- OpenAI
- AWS S3
- Python
- TypeScript
Turbopuffer alternatives
FAQs
How does it keep costs low?
It keeps data on object storage and caches hot data.
What search types are supported?
Vector and full-text.
Is it self-hosted?
It is primarily a managed service.
Work at Turbopuffer? Request updates to this profile. Update this profile
Compare with similar providers
All Data Infrastructure providersActiveloopMultimodal AI data lake and vector store (Deep Lake)
AerospikeReal-time NoSQL database built for low-latency, high-throughput workloadsAlgoliaSearch and discovery API platform with AI-powered retrieval
ApertureDataMultimodal database for AI with vector search and graph
ArangoGraph-native multi-model data platform for AI applications and agents
ArangoDBMulti-model database combining graph, document and search
ChromaOpen source AI-native vector database for embeddings and retrieval
CouchbaseDistributed NoSQL document database with cloud service and mobile sync