Weaviate

Weaviate

Weaviate is an open-source, cloud-native vector database that stores both objects and their vector representations. It enables semantic search, hybrid search (vector + BM25), Q&A, generative search, and multi-modal search at production scale — with modules for OpenAI, Cohere, Hugging Face, and more built-in.

Key Features

• Hybrid search: combine vector similarity and keyword (BM25) search with fusion ranking
• Native ML modules: integrate OpenAI, Cohere, Hugging Face, Google PaLM directly
• HNSW indexing with optional flat quantization for speed/memory tradeoffs
• Multi-tenancy: thousands of isolated tenants in a single deployment
• GraphQL + REST + gRPC APIs
• Kubernetes-native: Helm chart, horizontal scaling, replication

Quick Start

import weaviate

client = weaviate.connect_to_local()

collection = client.collections.create(
    name="Articles",
    vectorizer_config=weaviate.classes.config.Configure.Vectorizer.text2vec_openai()
)

collection.data.insert({"title": "AI in 2026", "body": "..."})

results = collection.query.near_text(query="machine learning trends", limit=5)

Install via ai-supply

npx ai-supply add weaviate-vector-database

Curated mirror of the open-source Weaviate (BSD-3-Clause). Get it from the source.