Models

The Models page, under Administration → Configuration → Models, is where you connect AI providers so their models can be used through WitnessAI, including in the Secure AI Portal. Once a provider is connected, its models can be selected elsewhere in the product, and WitnessAI applies your policies to every prompt and response sent to them.

Concepts

Provider: a service that hosts models, such as a public AI vendor or your own self-hosted endpoint.
Model: an individual model offered by a provider. After you connect a provider, you enable the specific models you want to make available.
Private provider: a self-hosted or internal provider that WitnessAI reaches at a custom Base URL rather than the vendor's public endpoint.

Prerequisites

The Admin role or higher.
An API key for the provider you want to connect.
For a private provider, the Base URL of the endpoint WitnessAI should connect to.

Add a provider

Select Add Provider to open the Add AI provider form.
Select the provider you want to add from the AI provider drop-down.
Select the Models you want to enable for that provider. This field is disabled until you choose a provider.
Complete the provider details:

Name: a label for this provider connection.
Private: turn this on if the provider is self-hosted or internal.
Base URL (optional): shown for private providers; the endpoint WitnessAI connects to.
API Key: the credential used to authenticate with the provider.

Select Configure provider to save.

The Add AI provider form, with AI provider and Models drop-downs, Name, Private, and API Key fields.

Add a private (self-hosted) provider

Select Add Provider.
Select the provider you want to add, then select its Models.
Turn on Private.
Enter a Name for the connection.
Enter the Base URL of your self-hosted endpoint.
Enter the API Key that your endpoint requires.
Select Configure provider to save.

Enable models

After a provider is connected, enable the individual models you want available through WitnessAI. Only enabled models appear for selection in the Secure AI Portal and elsewhere.

Verify

After saving a provider, a Provider added successfully confirmation appears. The provider then appears in the listing with its name in the Provider Name column and the enabled models listed in the Models column.

Settings reference

Field	Description
Name	A label for this provider connection.
Private	Marks the provider as self-hosted or internal.
Base URL (optional)	The endpoint WitnessAI connects to. Shown for private providers.
API Key	The credential used to authenticate with the provider.

Models Concepts Prerequisites Add a provider Add a private (self-hosted) provider Enable models Verify Settings reference