Skip to content

CommandLyne Guides

Step-by-step resources to help you deploy securely, govern your AI usage, and build automated workflows across your organization.

Sessions and Memory

Configure session limits and short-term, long-term and deep memory, with how retrieval works and the scope each memory layer actually covers.

Overview

Sessions and Memory control how the agent manages conversation history, context retention, and long-term knowledge.

Sessions

Sessions define how long conversations remain active and how they are grouped.

You can configure:

    • Max Age
    • Session Scope
    • Max Messages
    • Max Size

Default Max Age from the guide:

  1. 168 hours / 7 days

Memory Architecture

CommandLyne uses a three-layer memory structure.

  1. 1

    Short-Term Memory

    Stores recent interactions and supports immediate conversation continuity.

  2. 2

    Long-Term Memory

    Stores summarized daily knowledge for longer-term recall.

  3. 3

    Deep Memory

    Stores detailed memory records with metadata such as:

    • Dates
    • Tasks
    • Events
    • Context details

    Retention period from the guide:

    1 month

    After one month, deep memory is summarized and transferred into long-term memory.

Memory Retrieval Flow

The agent checks:

    • Short-Term Memory
    • Long-Term Memory
    • Deep Memory, if deeper detail is needed

Memory Scope

You can define memory scope, such as:

    • self
    • tree
    • agent
    • all

Need help setting up?

Get hands-on support for your deployment. Contact us for a guided onboarding session.