The Storage module provides comprehensive vector database support with multiple backend options, enabling you to choose the right storage solution for your performance and scalability requirements.

​
Core Storage Components

Foundation components for vector storage management:

Vector Database

Base vector database interface and core functionality

​
Vector Database Backends

Choose from multiple production-ready vector database backends:

Redis Vector Database

High-performance Redis-based vector storage

MongoDB Vector Database

MongoDB vector search capabilities

Qdrant Vector Database

Specialized Qdrant vector database integration

In Memory Vector Database

Fast in-memory storage for development and testing

TopK Vector Database

Optimized top-K retrieval vector database

​
Key Features

Storage components provide:
  • Multiple Backends: Support for various vector database technologies
  • Performance Optimization: Choose the right backend for your performance needs
  • Scalability: From in-memory testing to distributed production storage
  • Compatibility: Unified interface across all storage backends
  • Production Ready: Battle-tested integrations with popular vector databases
Each vector database backend is optimized for different use cases. Redis offers high performance, MongoDB provides rich querying, Qdrant specializes in vector search, and in-memory storage enables rapid development.

​
Backend Selection Guide

Choose the right vector database for your needs:
  • Redis: Best for high-performance scenarios with frequent updates
  • MongoDB: Ideal when you need rich metadata querying alongside vector search
  • Qdrant: Optimized specifically for vector similarity search operations
  • In-Memory: Perfect for development, testing, and small datasets
  • TopK: Specialized for scenarios requiring only top-K nearest neighbor results
Start with in-memory storage during development for fast iteration, then choose a production backend based on your specific performance and scalability requirements.

​
Storage Architecture

Storage components handle:
  1. Vector Indexing: Efficient organization of high-dimensional vectors
  2. Similarity Search: Fast nearest neighbor and similarity queries
  3. Metadata Management: Storage and retrieval of associated metadata
  4. Performance Optimization: Index tuning and query optimization