L2M

Appearance

Vector Store Nodes ​

Implemented vector-store nodes:

  • azure_ai_search_vector_store
  • qdrant_vector_store

Qdrant Vector Store ​

Node id: qdrant_vector_store

Supported actions:

  • get_ranked_documents
  • add_documents
  • retrieve_for_chain_tool
  • retrieve_for_ai_agent_tool

Common config:

  • endpoint (for example http://localhost:6333)
  • collectionName
  • secretRef.secretId (API key secret, optional for local unsecured Qdrant)
  • apiKeyHeaderName (default api-key)
  • useDemoFallback

Retrieval config:

  • queryText and/or queryVectorJson
  • filterJson (optional)
  • topK
  • contentField / metadataField

Mutation config:

  • documentsJson (JSON array of documents/points)

RAG Retrieve integration ​

RAG Retrieve supports Qdrant via:

  • vectorStoreId = qdrant-vector-store
  • credentials via embeddingSecretRef.secretId (or QDRANT_API_KEY)
  • connection config in vectorStoreConfig:
    • endpoint
    • collectionName
    • apiKeyHeaderName
    • contentField
    • metadataField
    • filter or filterJson