Weaviate
Free tierThe open-source AI-native vector database built for production-scale search, RAG, and agents
Free tier available·All audiences·API available·Open source
Key strengths
Billion-scale vector search with multi-tenancy supportBuilt-in embeddings — no external pipeline requiredDeployment-agnostic: cloud, self-hosted, or on-premUnified platform for search, RAG, agents, and memory (Engram)Production-hardened with SOC 2, HIPAA, RBAC, and high availability
Free tier + paid plans
Amsterdam, Netherlands
Founded 2019
Self-hostable
No ratings yet
- Vector similarity search — Store and query high-dimensional embeddings using HNSW indexing; supports
near_vector,near_text, andhybridquery modes with tunablealpha. - RAG pipelines — Combine Weaviate's retrieval layer with any LLM (OpenAI, Cohere, HuggingFace, etc.) for grounded, context-aware generation without external embedding pipelines.
- Agentic AI backends — Use Weaviate as the memory and retrieval layer for LLM agents; the Query Agent auto-translates natural language into optimized DB queries.
- Multi-tenant vector stores — Provision 50K+ isolated tenant indexes within a single cluster for SaaS platforms requiring strict data segmentation.
- Billion-scale production deployments — Handle workloads of 9B+ vectors in production with horizontal auto-scaling and cost optimization on Weaviate Cloud or self-hosted Kubernetes.
- Personalization & memory (Engram) — Persist and retrieve user-specific embeddings and interaction history to power adaptive, personalized AI experiences.