Agents Dashboard in the Build section of the Agentic Hub is where you create, view, and manage all the Agents within your Project.
In the upper-right corner of the screen, you'll find the + New Agent button. Clicking on this button offers three options:
-
Guides you step by step through creating a new Agent.
-
Allows you to manually configure all the details of a new Agent.
-
Allows you to import a LabPackage file containing an Agent and all the resources required for it to function (such as Integrations or other Agents included as resources). Selecting this option opens a panel where you can upload the file or drag and drop it.
When importing, keep the following in mind:
- The process creates the Agent and all its resources. If the Agent or any of its resources already exists in the Project under the same name or ID, the import will fail. Importing does not overwrite existing resources; it only adds new ones.
- The package must be a LabPackage generated using the Export Agent feature.
- Any external public resources (other Agents or Integrations) referenced in the package must already exist in the Project.
- Importing an Agent that was previously deleted from the same Project is not supported. The import process fails even if the Agent no longer appears in the Project.
By default, the Dashboard displays each Agent as a card.
Each card shows the Agent name, role, and a brief description, along with two status tags:
- Publication status (left tag): Indicates whether the Agent has been published to Solutions and its current state in the publication lifecycle. The possible values are:
- Not published: The Agent has not been published to Solutions.
- Published: The Agent is currently published and available to consumers in Solutions.
- Unpublished: The solution was published but is no longer accessible in the Solutions.
- Pending Review: The Agent has been submitted for review before publication.
- Pending Changes: The Agent has been published but has pending changes that have not yet been published.
- Build status (right tag): Indicates the current state of the Agent in the Build section. The possible values are:
- Draft: The Agent is still being configured and has not yet been saved. It is displayed with a dotted circle icon.
- Not Saved: The Agent was previously saved, but changes were made and those updates have not yet been saved. It is displayed with a dotted circle containing a check mark.
- Saved: The Agent has been saved with its latest configuration. It is displayed with a solid check mark inside a circle.
When the Agent has been saved at least once, a version badge appears beside this tag using the format V{version}.R{revision}. For example, V4.R75 Saved or V1.R3 Not saved. If the Agent has never been saved, only the Draft label is shown with no version badge.
Any changes you make to an Agent are automatically saved as a Not saved version. This ensures your progress is preserved even if you do not manually save the Agent.
You can also switch to a list view using the toggle icons in the upper-left area of the Dashboard. In this view, all Agents are displayed in a table with three columns: Agent Name, Agent Role, and Agent Purpose, along with the current build status at the end of each row.
Clicking an existing Agent opens a detailed panel on the right side of the screen. This panel displays additional information about the selected Agent, including its full description and AI model, among other details.
Each Agent card includes a ⋮ More Options menu. This menu allows you to perform additional actions, such as:
-
Configures whether a consumer can share, through a link, the conversation they are having with this Agent in the Run section (for example, None or Anonymous). For more information, see Sharing a conversation with an Agent with other consumers via a link.
-
Lets you define who can execute the Agent after it is published (for example, None or Project), according to the options described in Publishing an AI Solution.
-
Downloads a LabPackage containing:
- The Agent itself.
- Complete definitions of private resources (Integrations or Agents) and public resources from the same Project used by the Agent.
- References to external public resources used by the Agent.
This button is disabled until the Agent has been saved at least once. If there are unsaved changes, the export will include only the last saved version.
-
Opens the Generate client code panel, where you can copy a ready-to-use code snippet to test the Agent from your own environment. The panel provides two language options:
- cURL: Generates a command-line request including the required headers (Content-Type, project-id, and Authorization) and a JSON body containing the Agent model ID, a sample message, and the stream parameter.
- Python: Generates an equivalent script using the requests library, with separate headers and payload dictionaries, and basic response handling.
-
Permanently removes the Agent from the Dashboard.
-
Publishes the Agent to Solutions so consumers can discover it in the catalog and run it in the Run section, according to the configured execution permissions.
Each Agent card also includes a Test button and an Edit button. Use Test to open the panel on the right and interact with the Agent. Clicking Edit opens a new tab displaying the Agent name and description at the top.
A new tab also opens when you select New Agent Assisted, New Agent Manually, or Import Agent from the New Agent button. When creating an Agent using New Agent Manually, the tab opens as Untitled Agent.
On the left, you'll find the chat with Iris, the integrated Agent. You can interact with Iris to create or update an Agent, ask questions about the platform, and get assistance related to the documentation.
On the right, the following main tabs are displayed:
In the upper-right corner of the Agent editor, a three-dot (⋮) menu is available. This menu appears across all three tabs and includes the same options as the More Options menu available from each Agent card (except for Publish), with one additional option:
-
Opens the Version History panel on the right side of the Agent editor. This panel displays all saved versions of the Agent in chronological order, from newest to oldest. Each version shows its creation date and the user who saved it. The current version is visually identified.
Some versions may include one or more revisions. When available, expanding a version displays its associated revisions, including the revision number, creation date, and author.
You can use the search box at the top of the panel to find a specific version or revision.
Each version and revision includes a three-dot (⋮) menu with the following actions:
- Restore: Creates a new editable draft based on the selected version or revision without replacing the current version until it is saved.
- View Details: Displays the details of the selected version or revision in a read-only panel.
- Edit Details: Allows you to modify the metadata of the selected version or revision.
- Export: Downloads a LabPackage containing the selected version or revision.
