Skip to content
For the complete documentation index, see llms.txt. This page is also available as Markdown.

Vector stores

Set vectorStore on the client, or pass it on addDocument, updateDocument, deleteDocument, getDocument, listDocuments, and search. Leave it unset to use Postgres.

notifier.serverConfig = {
  vectorStore: { provider: "postgres" },
};

vectorStoreConfig is another name for the same object.

Store

provider

What you provide

Postgres

postgres

Nothing. This is the default. Keyword and hybrid search are available here.

Pinecone

pinecone

apiKey, and host or index. namespace is optional.

Weaviate

weaviate

host. apiKey is optional. className defaults to RagChunk.

Qdrant

qdrant

host. apiKey is optional. collection defaults to rag_chunks.

Azure AI Search

azure_search

endpoint and apiKey. index defaults to rag-chunks. apiVersion defaults to 2024-07-01.

S3 Vectors

s3_vectors

accessKeyId and secretAccessKey. Also indexArn, or both vectorBucket and index. region defaults to us-east-1. sessionToken is optional.

azure and azure_ai_search are accepted in place of azure_search. aws, aws_s3, s3, s3vectors, and amazon_s3_vectors are accepted in place of s3_vectors.

A vectorStore on one call overrides the client setting for that call. The embedding model stays as you set it.

→ How to set a vector store · Manage documents · Semantic search

Last updated