Purpose-built for AI embeddings.
High-performance vector database for similarity search, RAG, and AI applications. Store billions of vectors with sub-millisecond search latency.
Billions
Vectors
< 1 ms
Search
Up to 64K
Dimensions
HNSW / IVF
Indexes
AI-native search.
Billion-scale vectors with sub-ms search.
Billion-scale search
HNSW and IVF indexes for sub-millisecond similarity search across billions of vectors.
Hybrid search
Combine vector similarity with metadata filters. Structured + unstructured in one query.
Multi-tenancy
Namespace isolation for multi-tenant applications. Per-namespace access control.
Real-time upserts
Add, update, and delete vectors in real-time. No batch imports required.
Multi-region
Replicate collections across regions for low-latency global access.
Embedding pipelines
Built-in embedding generation from text, images, and audio.
Getting started
Launch your first instance in three steps. CLI, console, or API — your choice.
ur db vectordb create my-vectors \
--dimensions=1536 --metric=cosineVector search patterns.
RAG, recommendations, and semantic search.
Retrieval-Augmented Generation
Ground LLM responses with relevant context from your knowledge base.
View tutorialSuggested configuration
HNSW · Hybrid search · Real-time
Estimate your costs
Create detailed configurations to see exactly how much your architecture will cost. Pay for what you use, down to the second.
Configuration 1
Vector Database
Node Resources
Data & Backups
Plan & Commitment
Cost details
Sub-millisecond similarity search. RAG-optimized.
Works seamlessly with
Frequently asked questions
AI-native search.
Billion-scale vector search with sub-millisecond latency.