--- title: "How to Organize Objects with Groups, Keywords, Purposes, and Flags" description: "Classify registered objects using Groups, Keywords, Permissible Purposes, and Flags." ---
TL;DR
Four label types under Settings classify registered objects: Groups bucket objects by type, Keywords tag them for search, Permissible Purposes record what an object may be used for, and Flags mark a state with an audit trail. The first three also drive role Collection filters, so define them here first.
:::note[Admin guide] Each label has its own page under **Settings**, and you need write access to it. An item that's already in use shows a **Usage Count** and has some fields locked. ::: :::tip Groups, Keywords, and Permissible Purposes are the attributes you filter on when scoping a [role's](/user-guide/admin-guide/configure-roles/) access with Collection filters. Define them here to use them there. ::: ## Groups A group buckets objects of one type into categories (like `Auto Lending` or `Cards`) and sets the order they appear in. Groups are per object type, so a group made for models applies only to models, and there's no nesting. Go to **Settings → Groups** and click **+ New Group**: * **Object Type** * - which object type the group categorizes. Locked once the group is in use. * **Sort Order** * - display order in the UI (minimum 1). * **Description** * - what the group contains. Once created, the group is selectable when someone registers an object of that type, and the registry lists objects grouped by it. ## Keywords A keyword is a free-form tag that makes an object easier to find. It carries no state or access meaning; it just feeds search. Go to **Settings → Keywords** and click **+ New Keyword**: * **Name** * - the tag value. * **Type** * - the keyword type (Registry). * **Description** * - what the keyword means. Once created, a keyword can be attached during registration and is then searchable across the registry. ## Permissible Purposes A permissible purpose defines where an object is allowed to be used. Users assign one or more when registering an object, and the platform checks them across the object's lineage: if any of its dependencies carry a conflicting purpose, the object shows a **Permissible Purpose Violation** in its lineage view. Go to **Settings → Permissible Purposes** and click **+ New Permissible Purpose**: * **Name** * - the compliance purpose, such as `Fair Lending`. * **Description** * - the regulatory basis and scope of use. ## Flags A flag marks a state on an object and keeps an audit trail of every change. Use one for something you need to track over time, like a model that's due for review. Anyone with write access to Settings can create a flag; anyone in a workflow that oversees that object type can raise or drop it. Go to **Settings → Flags** and click **+ New Flag**, then give it a **Name** and **Description**. To raise a flag on an object: 1. Open the object and go to its **Flags** tab. 2. Click **Add Flag**, select the flag, and enter a **Description**. 3. Click **Submit Changes**. The flag's icon shows next to the object's name. Raising or dropping it requires a comment, and the platform records who changed it and when. ## What's next * Use these labels in [role Collection filters](/user-guide/admin-guide/configure-roles/) to scope access. * They appear as selectable metadata when users [register objects](/user-guide/how-tos/register-a-model-definition/).