Storage Overview

MemexLLM provides multiple storage backends for persisting conversation history. Choose the storage that best fits your needs or create your own custom storage.

Available Storage Backends

Memory Storage

• In-memory storage for development and testing
• No persistence across application restarts
• Fast and simple
• Learn more about Memory Storage

SQLite Storage

• File-based storage using SQLite
• Persistent across application restarts
• Great for single-machine applications
• Learn more about SQLite Storage

Coming Soon

We're working on additional storage backends:

PostgreSQL Storage

• Full SQL database support
• Great for distributed applications

Redis Storage

• Fast in-memory storage with persistence
• Great for high-performance applications

MongoDB Storage

• Document-based storage
• Flexible schema for complex metadata

S3 Storage

• Cloud storage support
• Great for serverless applications

Storage Features

All storage backends in MemexLLM provide:

• Thread Management: Create, read, update, and delete conversation threads
• Message Storage: Store and retrieve messages within threads
• Metadata Support: Attach custom metadata to threads and messages
• Message Limiting: Control the number of messages stored per thread
• Type Safety: Full type hints support

Custom Storage

Need a different storage backend? You can create your own:

• Implement the storage interface
• Full control over persistence and retrieval
• Learn how to create custom storage

Next Steps

• Try Memory Storage for development
• Use SQLite Storage for production
• Learn to create custom storage
• See example applications