Weaviate

Weaviate is an open-source vector database that stores objects and vectors and combines vector search with structured filtering, RAG and reranking. The server is written in Go (GitHub API and README).

Repository snapshot (GitHub API, 2026-11-19)

ItemDetail
Repoweaviate/weaviate, Go
LicenceSplit. The LICENSE file says code outside the wl directory is BSD-3-Clause; code inside wl is copyright Weaviate BV and available only under a separate enterprise licence. GitHub therefore reports no single licence.
Created2016-03-30
Activitylast push 2026-11-19; latest release v1.40.6 (2026-11-19)
Size16,957 stars, 1,439 forks
MaintainerWeaviate (company)

Capabilities (weaviate.io)

  • Hybrid search combining vector and keyword search with an alpha weighting; near-text and near-vector queries.
  • Built-in vectorisation through model providers such as OpenAI, Cohere and Hugging Face.
  • A Query Agent that turns natural-language questions into queries.
  • Multi-tenancy within one cluster; role-based access control, replication and high availability; the site names SOC 2 and HIPAA.
  • Clients for Python, Go, TypeScript and JavaScript, with GraphQL and REST APIs.

Deployment and billing

Docker, Kubernetes, AWS and GCP marketplace installs, and the managed Weaviate Cloud (Shared Cloud and Dedicated Cloud) with a free tier, usage-based plans and enterprise options. Rates are on the vendor’s pricing page. Index types, quantisation methods, sparse vectors and an embedded mode were not stated on the pages read.

Where it fits

Weaviate leans on integrated vectorisation and an agent layer, where qdrant emphasises filtering and quantisation and milvus emphasises scale. See vector-databases-compared.

Sources

  • weaviate.io, the Weaviate README, LICENSE file and GitHub API (2026-11-19)