DocsDatabases

Weaviate

Open-source vector database with strong hybrid search, modular embedders and a GraphQL-flavored API.

At a glance

License

BSD-3 (open source)

Hosting

Self-host or Weaviate Cloud

Default index

HNSW

Best for

Hybrid search + multi-tenancy

Define a collection

import weaviate from "weaviate-client";

const client = await weaviate.connectToLocal();

await client.collections.create({
  name: "Article",
  vectorizers: weaviate.configure.vectorizer.text2VecOpenAI(),
  properties: [
    { name: "title",   dataType: "text" },
    { name: "content", dataType: "text" },
  ],
});

Hybrid query

const articles = client.collections.get("Article");

const result = await articles.query.hybrid("vector indexes for RAG", {
  alpha: 0.5,   // 0 = pure BM25, 1 = pure vector
  limit: 10,
});

Weaviate ships with built-in vectorizer modules — you can pass raw text and let the server call the embedding model for you. Convenient for prototypes, but most teams disable it in production to keep control of embedding versions.