Skip to content

Containers and Scopes in XTM One

Some pages in XTM One are about one specific object. Others are about a container or a scope that groups many objects together.

What a container means here

In XTM One, a container is any object or view whose main role is to hold, group, or organize related work.

A scope is the boundary that tells you where something is visible, reusable, or operationally relevant.

Main container types

Conversations

A conversation contains the messages, context, and chat history for one interaction thread.

Use it when you want to keep one discussion together over time.

Flows

A flow groups multiple agents into a coordinated workflow.

Use it when you want to understand a process rather than a single assistant.

Knowledge bases

A knowledge base groups documents and import sources into one searchable corpus.

Use it when you want agents to rely on the same body of information.

AI Catalog entries

The AI Catalog groups shareable resources into published packages that other users can review and clone.

Use it when you want distribution, not only local configuration.

Main scope types

Personal scope

A personal object belongs to one user.

This is the safest place for experimentation because your edits stay local.

Group-shared scope

A shared object is visible and editable by the members of the selected group.

This is a collaboration scope, not a read-only publication scope.

Company-managed scope

A company-managed object is visible to everyone, but only admins usually edit it.

This is the broadest scope in normal day-to-day use.

How containers behave

Containers usually do one or more of these things:

  • show the items they contain
  • let you add or remove linked items
  • expose an aggregate activity view
  • provide a focused view of one operational context

For example:

  • a flow shows the agents inside it and their shared activity
  • a knowledge base shows the documents and sources inside it
  • a conversation shows the messages inside it

How to read scope correctly

Before changing something, ask:

  • am I editing the container itself
  • am I editing an item inside the container
  • will this change affect only me, my group, or the whole company

Those three questions prevent most accidental changes.

What containers do not mean

Being inside a container does not always mean ownership.

For example:

  • an agent inside a flow is still its own object
  • a prompt attached to an agent is still a prompt that may be reused elsewhere
  • a catalog item added to your account usually becomes a new local copy, not a live link to the source

Practical rule

If a page helps you understand a collection, a boundary, or a workspace, you are probably looking at a container or a scope.

If a page is about one reusable thing with its own identity, you are probably looking at a standalone object.

Next step

Continue with Reuse and deduplication, which explains when XTM One keeps a shared object, when it creates a copy, and how to avoid duplicate setup.