By default, Threado uses the OpenAI GPT-4 model to generate answers. For most of the customers, this default option is the best and they don't need to change anything here. However, if you have specific requirements around data security or you want to experiment with your own custom models, you can add your own following the given instructions.

Navigate to the AI settings tab on the Settings dashboard.

The first option allows you to connect your own Generative AI model.

Click on + Own Model and then select from the drop-down which model you'd like to connect.

There are two models to choose from -

Choose an Open AI model from the drop-down. This will prompt you to add the API key for that model.

Navigate to https://platform.openai.com/ to access your API key.

After you've logged in, go to the API Keys tab. Here, you can create a new Secret key and copy the credentials.

Paste the credentials on Threado AI and connect to add your model.

Choose an Azure model from the drop-down. This will prompt to establish the connection by adding an API Endpoint, Model Deployment Name, and API key.

Navigate to your Azure developer account and log in. Search for 'Azure OpenAI' in the search bar and open that portal.

Select the model you want to connect to and navigate to the Develop tab.

Now, copy the API Endpoint and API key from here.

Now, navigate to the Go to Azure OpenAI Studio tab at the top.

To get the Deployment name, go to the Deployments tab and copy the name of the deployed model.

After adding these details, click Connect to verify them and upon successful connection, you will see an option to choose that model on the main dashboard.

Select this model to start using it.

In terms of the process of connecting an Embedding AI model, is similar to connecting a Generative AI model.

From the drop-down select which model you would like to connect.

There are two models to choose from -