This article explains how to securely store your OpenAI credentials in Zammo and configure and manage OpenAI GPT and Embeddings models. It also covers how to add new model configurations, test connections, and set active models for use in your environment.

To set your OpenAI model version, URL and key:

1. Select the AI Models option in the Developers tab in the Portal.

1. Here you’ll see the models listed for Chat and Embeddings along with their deployment names. You can use the auto-deployed models.

2. To add a new model, select the Add new OpenAI model configuration option**.**

Select either the GPT or Embeddings model type.

In the modal that follows, enter your information for the following fields:

• Name - The name of the model configuration. This will help identify the model configuration in your list.
• Description - A brief description of the model to clarify its purpose.
• OpenAI URL - The endpoint URL for your OpenAI connection.
• OpenAI Secret - The secret key for authenticating with OpenAI.
• GPT/Embeddings Deployment Name - The deployment name for the GPT/Embeddings model.

<aside>

If you need help locating your Azure OpenAI URL and key, please see the instructions provided in the article Evaluate Zammo.

</aside>

Select Test to test the connection and detect the model version.

When the test is successful, a green notification banner will display at the bottom of the modal and the model version detected will be inserted into the disabled Model Version field.

Choose Save.

In the Custom AI Model Management page, the new OpenAI model is listed. Choose the Set Active option to select the active model.