Memory Stores

Memory stores provide persistent storage capabilities for AI agents, enabling them to retain and retrieve information across conversations and sessions.

What is a Memory Store?

A Memory Store is a structured storage system designed to give AI agents the ability to remember information over time. It acts as a persistent knowledge base that agents can reference, update, and learn from, across multiple interactions.

Unlike temporary conversation context, Memory Stores provide durable storage that persists indefinitely, allowing agents to build up knowledge and maintain context across extended periods.

Long-Term Memory

Currently, we only support long-term memory storage. This means:

  • Information stored in memory persists indefinitely.
  • Agents can access historical data from previous interactions.
  • Memory remains available across different sessions and contexts.
  • No automatic expiration or cleanup of stored information.

Long-term memory is essential for building sophisticated agent applications that need to maintain state, track progress, or accumulate knowledge over time.

Importance of Memory for Agents

Memory Stores are crucial for creating intelligent and context-aware agents

Contextual Continuity

Agents can maintain context across multiple conversations, remembering previous interactions, preferences, and decisions made in earlier sessions.

Knowledge Accumulation

As agents interact with users and process information, they can build up a knowledge base specific to their domain or use case, becoming more effective over time.

Personalization

Memory enables agents to remember user-specific information, preferences, and history, allowing for highly personalized experiences.

Complex Task Management

For multi-step processes or long-running tasks, memory stores allow agents to track progress, maintain state, and resume work seamlessly.

Architecture

The memory store system follows a hierarchical organization


graph TD
    A[Memory Stores] --> B[Memory 1]
    A --> C[Memory 2]
    
    B --> E[Document 1.1]
    B --> F[Document 1.2]
    B --> G[Document 1.3]
    
    C --> H[Document 2.1]
    C --> I[Document 2.2]
    C --> J[Document 2.3]
    
    classDef storeClass fill:#0F172A,stroke:#3B82F6,stroke-width:3px,color:#FFFFFF
    classDef memoryClass fill:#1E293B,stroke:#8B5CF6,stroke-width:2px,color:#FFFFFF
    classDef documentClass fill:#334155,stroke:#10B981,stroke-width:2px,color:#FFFFFF
    
    class A storeClass
    class B,C memoryClass
    class E,F,G,H,I,J documentClass

Memory Stores: The top-level container that organizes all memory for your application or use cases

Memory: Individual memory units that contain specific information or knowledge about an entity

Memory Documents: The actual content items stored within memories, containing the data agents need to remember

👉

To get started, see Using Memory Stores.



What’s Next