Skip to content
Multi-Agent Orchestrator

Storage overview

The Multi-Agent Orchestrator System offers flexible storage options for maintaining conversation history. This allows the system to preserve context across multiple interactions and enables agents to provide more coherent and contextually relevant responses.

Key Concepts

  • Each conversation is uniquely identified by a combination of userId, sessionId, and agentId.
  • The storage system saves both user messages and assistant responses.
  • Different storage backends are supported through the ConversationStorage interface.

Available Storage Options

  1. In-Memory Storage:

    • Ideal for development, testing, or scenarios where persistence isn’t required.
    • Quick and efficient for short-lived sessions.

  2. DynamoDB Storage:

    • Provides persistent storage for production environments.
    • Allows for scalable and durable conversation history storage.

  3. SQL Storage:

    • Offers persistent storage using SQLite or Turso databases.
    • When you need local-first development with remote deployment options

  4. Custom Storage Solutions:

    • The system allows for implementation of custom storage options to meet specific needs.

Choosing the Right Storage Option

  • Use In-Memory Storage for development, testing, or when persistence between application restarts is not necessary.
  • Choose DynamoDB Storage for production environments where conversation history needs to be preserved long-term or across multiple instances of your application.
  • Consider SQL Storage for a balance between simplicity and scalability, supporting both local and remote databases.
  • Implement a custom storage solution if you have specific requirements not met by the provided options.

Next Steps

By leveraging these storage options, you can ensure that your Multi-Agent Orchestrator System maintains the necessary context for coherent and effective conversations across various use cases and deployment scenarios.