Weaviate
Weaviate is an open-source vector database that stores objects alongside their vector embeddings, combining approximate nearest-neighbor vector search with structured filtering and ordinary CRUD semantics. It is commonly used as the retrieval layer behind semantic search and retrieval-augmented generation applications.
Baseline topologies
| Topology | Minimum nodes | Roles |
|---|---|---|
| standalone-hybrid-search | 1 | weaviate |
| replicated-search-cluster | 2 | weaviate |
What drives the cost
- Vector count (millions)
- Approximate number of stored vectors before replica or multi-tenant overhead.
- Embedding dimension
- Dimension of the embeddings stored in the primary collections.
- Peak query concurrency
- Expected peak concurrent semantic or hybrid-search requests.
Where it runs
bare_metal, dedicated, cloud_vm
Size Weaviate and get a cost estimate