Data & Retrieval
Vector databases
Managed storage and similarity search for embeddings.
What it is
A vector database stores embeddings and finds the nearest ones to a query vector in milliseconds, often combined with keyword search and metadata filters. It is the retrieval layer of RAG and semantic search.
How pricing works
Mix of storage (per GB or per vector), read/write units and sometimes pods. We normalise to 1M 1024-dimension vectors with light query traffic for a month.
We compare everything in one unit: USD per month for 1M vectors (1024-dim).
Typical uses
- RAG knowledge bases
- Semantic and hybrid search
- Recommendation and deduplication
What to compare
- Query latency at your scale
- Hybrid (keyword + vector) search
- Metadata filtering
- Serverless vs provisioned
Watch out for
- Query-unit pricing can explode with heavy traffic; model your read volume.
All 7 APIs, cheapest first
Free to browseTick 2 to 4 APIs to compare them side by side.
| API | $ / mo · 1M vectors | Serverless | Hybrid search | Open source | Quality | Price data | Compare |
|---|---|---|---|---|---|---|---|
| turbopuffer turbopuffer | $4.00 | Yes | Yes | No | 80 | Unverified sample | |
| Chroma Cloud Chroma Cloud | $5.00 | Yes | Yes | Yes | – | Unverified sample | |
| Zilliz Cloud Serverless Zilliz Cloud (Milvus) | $10.00 | Yes | Yes | Yes | – | Unverified sample | |
| Pinecone Serverless Pinecone | $15.00 | Yes | Yes | No | 85 | Unverified sample | |
| Qdrant Cloud Qdrant Cloud | $25.00 | No | Yes | Yes | 84 | Unverified sample | |
| Weaviate Cloud (Flex) Weaviate Cloud | $45.00 | Yes | Yes | Yes | – | Unverified sample | |
| MongoDB Atlas Vector Search MongoDB Atlas Vector Search | $60.00 | Yes | Yes | No | – | Unverified sample |