Weaviate

Open-source vector database with hybrid search built in

AI Models & PlatformsFree planOverseasβ˜…β˜…β˜…β˜…β˜† 4.0

What is Weaviate?

Weaviate is an open-source vector database that stores objects and their embeddings, then supports semantic, keyword and hybrid search over them. It can run in Docker on your own machine, in Kubernetes, or as Weaviate Cloud, so the same APIs work whether you self-host or buy the managed service. Modules handle vectorisation for you, calling embedding providers directly so you can insert raw text instead of computing vectors yourself. The differentiator is hybrid retrieval: combining keyword scoring with vector similarity often beats either method alone on real queries.

Last updated: 2026-09-20. This site only provides an index; for exact features, pricing, and licensing, see the official website.

Key features

  • Hybrid search blending vector similarity with keyword scoring
  • Vectoriser modules for OpenAI, Cohere, Hugging Face and others
  • GraphQL and REST APIs plus clients for several languages
  • Multi-tenancy support for per-customer data isolation
  • Generative search module for retrieval-augmented answers
  • Self-hosted Docker deployment or managed cloud service

Pros & cons

Strengths

  • Open source, so self-hosting carries no licence cost
  • Hybrid search improves recall on keyword-heavy queries
  • Same API whether you run it locally or in the cloud

Watch out for

  • Self-hosting means you own scaling and backups
  • Resource use is heavier than a minimal vector index
  • Module configuration adds concepts to learn up front

Best for & use cases

hybrid search, rag retrieval, self-hosted vector search, multi-tenant knowledge bases

If you're comparing similar products, check the alternatives below, or browse all tools in the AI Models & Platforms category.

FAQ

What makes hybrid search useful?

Pure vector search misses exact terms like product codes and names, while keyword search misses paraphrases. Hybrid search runs both and merges the scores, which usually improves recall on real queries.

Can I run it for free?

Yes. Weaviate is open source and can be self-hosted with Docker or Kubernetes at no licence cost, so you only pay for the machines it runs on. The managed cloud has a free sandbox tier.

Do I have to compute embeddings myself?

Not always. Vectoriser modules call embedding providers for you, so you insert raw text and Weaviate handles vectorisation. You can also supply your own vectors if you already generate them.