Install and Register the Gaia AI Engine

Deployment Model: Self-Managed

29 January 2026

To install and configure the Gaia AI Engine:

Log in to Helios Self-Managed.
Click the app-selector menu and navigate to Insights > Gaia.

The Gaia onboarding page is displayed:
Install the tar file that you downloaded. You can do one of the following:
- Download the tar file directly onto a node within the OpenShift cluster.
- Download the tar file onto another machine that has both internet access and connectivity to the OpenShift cluster. For more information, see Jump Host.
Extract the tar archive and navigate to the gaia-ai-engine directory. This directory contains the necessary installation scripts and configuration files required to deploy the Gaia AI Engine.
Click Register and on the Register Gaia’s AI Engine page, enter the following details and click Download:
- Gaia AI Engine Name: Enter the name of the Gaia AI Engine.
  
  This name is displayed in the Gaia Region drop-down list when creating a dataset.
- Domain Name: Enter the domain name.
- Port Number: Cohesity recommends that you use port number 443.
The configuration file is downloaded in JSON format.
- Transfer this JSON file to the gaia-ai-engine directory. The configuration file contains essential values and certificates required to establish the connection.
- Transfer the S3 credentials and S3 certificate, which are used to authenticate with the S3 endpoint, to the gaia-ai-engine directory.
To begin the installation, use one of the following commands:
- Single-node installation:
```
./gaia-ai-engine.sh --install --deployment-mode singlenode
```
- Cluster installation (default option):
```
./gaia-ai-engine.sh --install --deployment-mode cluster
```
Confirm acceptance of the End User License Agreement (EULA) when prompted.
Acknowledge that all prerequisite steps have been completed.
If your cluster does not have any GPUs, acknowledge the prompt and continue with the installation.
Enter the absolute path to the Gaia AI Engine JSON configuration file that you previously copied to the machine.

If the file path is incorrect, the script will notify you and you must restart the installation.
Enter the S3 endpoint name, S3 access key, and S3 secret key.
Enable SSL and provide the path to the certificate file you copied earlier.
Choose the storage provider class to provision persistent volumes in the Kubernetes cluster where Gaia AI Engine is deployed.
Enter the container registry address to be used for image storage.

If your registry requires authentication, create an image pull secret by providing the necessary credentials.

If no credentials are required, you may leave the username and password fields empty.

The installation process begins. Note the following:
- You may see warnings if there is no GPU available, but the installation will proceed.
- The script downloads the required images and pushes them to the specified container registry.
- The script first deploys all required Custom Resource Definitions (CRDs) and dependencies, followed by the main application.
After the installation is complete, click Verify Connection:

If the installation is successful, the status changes to Healthy.

You can view the connection status and Gaia AI Engine details in the Settings > AI Engine Management page:

If the installation is unsuccessful, the status changes to Failed. If the installation fails, contact Cohesity Support.
Click Create Dataset.

Manage AI Engine

The Settings > AI Engine Management page displays the connection status of the Gaia AI Engine. From this page, you can also edit the AI Engine details or delete the AI Engine. The following details are displayed:

AI Engine name
Connection status
Last updated date and time

Insights Documentation