Back to Insights
Artificial Intelligence•January 2, 2024•9 min read

Weaviate: Open Source Vector Database

Weaviate provides open-source vector search with hybrid search and automatic vectorization.

#weaviate#vector-database#search#embeddings

Weaviate combines vector and keyword search in one database. Built-in vectorization modules connect to embedding APIs. GraphQL interface enables flexible queries. Open-source with managed cloud option.

Schema Design

Define classes representing your data types. Configure vectorizers for automatic embedding. Set up property types for filtering. Plan sharding for large datasets.

  • Define schemas matching your data model
  • Configure appropriate vectorizer modules
  • Enable hybrid search combining vector and keyword
  • Use filters for metadata-based refinement
  • Monitor query latency and resource usage

Query Capabilities

Near text queries find semantically similar objects. Near vector queries use pre-computed embeddings. Hybrid search combines approaches. Filters refine results by properties.

Tags

weaviatevector-databasesearchembeddingsopen-source