Skip to main content

Reveal Review Publication

Environment setup

The Environment setup page allows you to accomplish any of these optional tasks:

  1. Third Party System Administrator Access (for entering Reveal AI from Relativity® as an admin).

  2. NexBERT Setup - Configure your Reveal AI environment for the use of NexBERT models.

  3. Set up Cache.

  4. Trusted URLs (for your Relativity® site).

  5. Keycloak Access Setup (as of Version 3.00, this is configured outside of Reveal AI across the Reveal environment).

1. Third Party System Administrator Access

You may use this section to allow users in the provided Relativity® group to sign into Reveal AI from Relativity® with a System Administrator account. To set up Third Party System Administrator Access perform the following:

  1. Enter Reveal AI as an admin directly (not by way of Relativity®).

  2. Go to SYSTEM > Environment setup > Third Party System Administrator Access:

  3. Check the checkbox for Relativity® access enabled.

  4. Designate a Relativity® Group Name that indicates System Admin status such as Reveal AI Sys admin.

  5. Hit Save.

  6. Enter Relativity®.

  7. Create the identical Relativity® Group Name in Relativity®.

  8. Add those Relativity® users whom you want to enjoy admin status in Reveal AI (including yourself) to this group.

  9. Sign into a storybook from Relativity®.

  10. Your Reveal AI system admin status verifies success.

2. NexBERT setup

NexBERT setup is only required if you are planning to use NexBERT models. Click on the pen icon next to the title of this Environment Setup section to edit. Depending on environment location, choose either Azure or On premise.

Configuring for Azure

If you choose Azure the fields below require population:


You must enter your Azure environment to acquire this information. This involves a series of steps.

First, enter your Azure environment. Go to Azure Active directory:


Choose App registrations and the following page appears:


Choose + New registration and the following page appears:


Enter the name of the new application (in this case: “reveal-demo”) and hit Register. This will provide you the Client ID (Application {client} ID in Azure) and the Tenant ID (Directory {tenant} ID in Azure) for the Reveal AI NexBERT setup form.


Now choose Certificates and secrets...


...and choose +New client secret. Fill out the “Description” and choose Add

You now see the new value for the Client secret.


Copy this value and paste into the NexBERT setup form NOW because you lose access to the client secret value when you leave this page.


The next step is to create a workspace. Go to AI + Machine Learning and then choose Machine Learning.


This takes you to the Machine Learning page:


...where you provide:

  • Workspace Name: Create a “Workspace name” and enter it as Workspace in the Reveal AI NexBERT setup form.

  • Resource Group: Choose the “Resource group” from the dropdown or create a new one and enter it as Resource group in the Reveal AI NexBERT setup form.

  • Region: Your workspace should be in the same location as your data storage. Choose the “Region” from the dropdown and enter it as Region in the Reveal AI NexBERT setup form.

  • Storage account, Key vault and Application insights are generated from the Workspace Name.

When these items are completed click Review and Create

The Machine Learning form appears for your review:


Choose Create. Creation will take a few minutes. You will see this when your workspace is created:


Hit on Go to resource and get the Subscription ID as shown in the image below. The “Subscription Id” should be entered as Subscription ID in the Reveal AI NexBERT setup form. 


Now to give the application access to the workspace you return to the Azure Home page and choose All resources in the left panel of Azure services and then choose the workspace (reveal-demo) from those listed under Name.



  • Choose Access control (IAM),

  • Add Role assignments,

  • Choose “Contributor” as the Role,

  • Select Next to assign to Members (User, Group or Service, or to a Managed Identity).

  • Choose the app we just created (“reveal-demo”), and then

  • Choose Select.

  • Click Review + assign in the bottom left corner of the screen when complete.

Next go to the Machine Learning Studio:


Choose Compute to open the Compute screen. This is where you add virtual machines that may be activated for intense processing cycles such as NexBERT scoring runs and then spun down when the process is complete to save empty processor cycle costs.


First choose Compute Cluster then + New.  The New compute cluster window appears.


The following items are needed:

  • Computer name: This must always be “default-cluster”.

  • Virtual machine tier: Must be Dedicated.

  • Virtual machine type: Switch this to “GPU (Graphics Processing Unit)”.

  • Virtual Machine Size: After choosing Select from recommended options choose Standard_NC6 from the list.

  • Minimum number of nodes: This must be 0 (that is zero).

  • Maximum number of nodes: This must be 6 (that is six).

  • Idle seconds before scale down: Populated automatically.

Click Create when complete.


Next, go to the Machine Learning studio and choose Datastores to open the Datastores page. It will appear with four default datastores, but you will want to add a new one for NexBERT scoring runs.

Choose + New datastores and the New datastore form appears:


Select From Azure Subscription to add the Subscription ID.

Choose Storage account and Blob container names from drop-down lists.

Respond Yes to Save credentials with the datastore for data access.

The Authentication Key will be needed to complete the Datastore authentication process. The Authentication Key is retrieved by returning to All resources.

Choose the application workspace created earlier (here, reveal-demo) and click on the link for the Storage Account.


In the Storage account screen select Access Keys. To show the keys available click Show keys at the top of the screen and copy one of the keys that appears.


Now return to the Datastore page and paste the key copy into the Account key box:


You now have the following input to copy from Azure and paste into the NexBERT setup form:

From Azure

Into the NexBERT setup form

“Datastore name”

“Datastore name”

“Storage account”

“Storage account name”

“Blob container”

“Container name”

“Account key”

“Storage account key”

The final item for the NexBERT setup form is the “Experiment name”. This requires command level input.

First go to the Azure services page and choose the cloud shell.


As shown above, login to azure shell and then switch to bash  

az login

Paste the code in the url that shows  

az extension add -n azure-cli-ml

Now you can run command to create a new experiment:  


Return to the Azure Machine Learning studio and choose Experiments.

The Experiments page will appear with the name of the newly created experiment.


The name of the experiment is copied and pasted into the “Experiment name” box of the NexBERT setup form. This concludes the process of filling out the NexBERT setup form.

To summarize:

From Azure

Into the NexBERT setup form

Application (client) ID

Client ID

Client secrets Value

Client secret



Subscription Id

Subscription ID

Directory (tenant) ID

Tenant ID

Workspace Name


Resource Group

Resource group


Experiment name

Datastore name

Datastore name

Storage account

Storage account name

Account key

Storage account key

Blob container

Container name

The NexBERT setup section of your Environment setup page now looks something like this:

Configuring for “On premise”

If your NexBERT setup will be in your own environment, you choose On premise and the fields below require population:


And enter the following required information:

  • Remote machine name*: Name the designated hardware.

  • Username*: Enter your username required to access the environment.

  • Password*: Enter the required password.

  • VM processing unit*: Choose either CPU or GPU.

  • Storage Path*: Provide the storage path for the environment.

Then choose Save.

For further information see Reveal AI Installation Guide.

3. Cache setup

On the Environment setup page underneath NexBERT setup is the Cache Setup section, where the caching method may be selected and additional password security set.


The choices for cache setup are:

  • In Process - Cache data in memory.

  • SQL Server Memory-Optimized Table - In order to use this option, the selected SQL database needs to be configured to add a new filegroup with MEMORY_OPTIMIZED_FILEGROUP option. Recommend using a new SQL database and using the following script to add the new file group.

    • Before running the script below, create a new SQL database.

    • Change the "filename" to point to a valid SQL path on your SQL server.

      use nexlp_cache

      ALTER DATABASE nexlp_cache ADD FILEGROUP nexlp_cache_filegp


      ALTER DATABASE nexlp_cache ADD FILE (



      TO FILEGROUP nexlp_cache_filegp;

  • Redis - Using Redis sever as the caching solution (default).

Enter the new password twice and choose Save. When the password is accepted, the page now shows The cache is ok. Password is set.

For more information, please refer to the Reveal AI Installation Guide, Section 13. Enhanced Caching Server Service Security.

4. Trusted URLs

You may use this section to provide an additional layer of security for your Relativity® connection by adding your Relativity® URLs.

Use the Actions tools to edit or remove.

5. Keycloak Access Setup

Keycloak supports fine-grained authorization policies and is able to combine different access control mechanisms such as OAuth. Keycloak can be configured and tested in the Keycloak access setup section of Environment setup.


This section has been removed from Reveal AI 3.00 as a part of the overall Reveal environment.