Packages
@anvia/pgvector: Overview
Postgres pgvector store adapter for Anvia retrieval and semantic search.
What it is
Postgres pgvector store adapter for Anvia retrieval and semantic search.
Use @anvia/pgvector when the application needs Postgres with pgvector as the backing store for embedded documents and semantic search. It is one of the storage adapters that expose a provider-neutral VectorSearchIndex.
Where it fits
@anvia/pgvector stores embedded documents in Postgres with pgvector and returns indexes that implement VectorSearchIndex. Agents can use the index as dynamic context or as a search tool without depending on the database client.
The package owns connection setup, document upsert, metadata filter translation through filterToPgVectorWhere, and search result mapping. Keep document loading, chunking, embedding model choice, tenant scoping, and ingestion jobs in application code.
Public surface
The main documented exports are PgVectorStoreConnectOptions, PgVectorStore, PgVectorIndex, filterToPgVectorWhere. The reference page lists the package entrypoint and public symbols that are checked by the docs reference coverage script.