Weaviate API
Weaviate API
Weaviate is an open-source vector database — supports vector search + keyword search + built-in vectorization (OpenAI/Cohere/HuggingFace and other models).
Open-source (Apache 2.0) + self-host
More complex ops than Pinecone (self-host requires cluster management)
Docker run weaviate for local instance, or sign up Cloud at weaviate.io. Python: client.collections.create("Article", vectorizer_config=...)
Uptime · 30-day window
About this API
Weaviate is a Dutch open-source vector database company founded 2019, one of the main open-source alternatives to Pinecone (others are Qdrant and Milvus). Differentiators: (1) Built-in vectorization — no need to call OpenAI embeddings before inserting vectors; Weaviate accepts raw text and auto-converts via configured model (OpenAI, Cohere, HuggingFace, etc.); (2) Native hybrid search — simultaneously uses BM25 keyword search and vector search with fused ranking (production RAG often outperforms pure vector with hybrid); (3) GraphQL interface for elegant complex queries (combining metadata filter + vector + keyword). Production scenarios are mainly self-host (Docker + cluster); cloud version Weaviate Cloud suits mid-size companies wanting open-source DB without ops.
What you can build
- 1Enterprise RAG (self-host, not cloud)
- 2Hybrid search (vector + keyword)
- 3Multi-tenant apps (flexible schema design)
- 4Open-source LLM full-stack
Strengths & limitations
Strengths
- Open-source (Apache 2.0) + self-host
- Built-in vectorization (no separate embedding API call)
- Native hybrid search support
- Flexible GraphQL queries
Limitations
- More complex ops than Pinecone (self-host requires cluster management)
- Production tuning requires vector DB experience
- Cloud version (Weaviate Cloud) relatively expensive
Example request
curl https://weaviate.io/<endpoint> \
-H "Authorization: Bearer $API_KEY"
# Some providers use X-Api-Key instead — verify in the docs.Getting started
Docker run weaviate for local instance, or sign up Cloud at weaviate.io. Python: client.collections.create("Article", vectorizer_config=...)
FAQ
Weaviate vs. Qdrant?+
Ecosystem maturity: Weaviate slightly ahead. Performance: Qdrant faster on some benchmarks. Hybrid search: Weaviate more native.
Self-host or Cloud?+
Startup / trial: Cloud. Production + cost-conscious: self-host (Docker + a few machines).
Technical details
- Auth type
- api_key
- Pricing
- freemium
- Rate limit
- 自托管无限；Weaviate Cloud 按 dimension 和 query 计费
- Protocols
- REST, GraphQL, gRPC
- SDKs
- python, javascript, typescript, go, java
- Response time
- 431 ms
- Last health check
- 5/12/2026, 7:38:37 AM