Feature Request: Direct Support for OpenSearch as a Vector Store

[Akhil-Pathivada]

🚀 The feature

Description

Mem0 currently lacks native support for OpenSearch as a vector store. Given OpenSearch's popularity as a scalable open-source search and analytics suite with k-NN vector search capabilities, adding direct support would benefit many users.

Use Case

We use OpenSearch k-NN 24 for vector storage and similarity search in our AI application. Integrating Mem0 would enhance our application with a memory layer for personalized and context-aware interactions, but the lack of direct OpenSearch support adds unnecessary complexity.

Proposed Solution

Implement an OpensearchVectorClient within Mem0, similar to the existing clients for other vector stores. This would allow Mem0 to directly interact with OpenSearch for storing and retrieving vector embeddings, streamlining the integration process.

Motivation, pitch

As organizations increasingly leverage Vector DBs for AI-powered applications 4, the need for seamless integration between memory layers like Mem0 and popular vector stores becomes critical. OpenSearch is a widely adopted open-source search and analytics suite that offers robust k-NN vector search capabilities, suitable for large datasets. By natively supporting OpenSearch, Mem0 can unlock significant benefits for users already invested in the OpenSearch ecosystem, enabling them to easily add a sophisticated memory layer to their AI applications.