Zep Collection - Cloud

​

Node Details

• Name: zepCloud

• Type: Zep

• Version: 2.0

• Category: Vector Stores

​

Base Classes

• Zep

• VectorStoreRetriever

• BaseRetriever

​

Credentials

• Type: JWT authentication

• Name: zepMemoryApi

• Description: Configure JWT authentication for your Zep instance (Optional)

​

Input Parameters

1. Document (optional)

   • Type: Document

   • List: true

   • Description: Input documents to be stored in the Zep collection

2. Zep Collection

   • Type: string

   • Placeholder: my-first-collection

   • Description: Name of the Zep collection to use

3. Zep Metadata Filter (optional)

   • Type: json

   • Description: JSON filter to apply on metadata when retrieving documents

4. Top K (optional)

   • Type: number

   • Default: 4

   • Description: Number of top results to fetch

5. MMR Parameters (optional)

   • Fetch K

   • MMR Lambda

   • Description: Parameters for Maximal Marginal Relevance search

​

Outputs

1. Zep Retriever

   • Type: Retriever

   • Base Classes: [Zep, VectorStoreRetriever, BaseRetriever]

2. Zep Vector Store

   • Type: Vector Store

   • Base Classes: [Zep, ZepVectorStore]

​

Functionality

1. Upsert:

   • Adds or updates documents in the Zep collection

   • Returns the number of added documents and the added documents themselves

2. Init:

   • Initializes the Zep vector store with the given configuration

   • Sets up metadata filters if provided

   • Returns either a vector store or a retriever based on the node data

​

Usage

This node is particularly useful for:

1. Storing and retrieving embedded data in a scalable manner

2. Performing similarity searches on stored data

3. Implementing MMR search for diverse result retrieval

4. Integrating Zep’s vector store capabilities into larger language model applications

​

Note

The node uses FakeEmbeddings for document storage, which means it expects pre-embedded documents. For production use, you should replace this with actual embedding functionality.