To set up GasbyAI with Azure OpenAI, you'll need to create a Resource (to get Resource name), create a model (to get deployment name*), and get the API Key. The following steps will help you get them from Azure OpenAI services

1. Create your Azure OpenAI resource

1. Login into your Azure Portal, click on Create a resource
2. Search for Azure OpenAI, and click Create at the bottom of the card. Then choose Azure OpenAI
3. On the Create Azure OpenAI page, fill in the details.

Once completed, your Azure resource is created. Note that you'll need the Resource name for provider settings

2. Get API Key

API Keys will be available for each resource. To get your API key

1. Navigate to the Azure Portal
2. Under Resources, select your new resource
3. On the Resource page, under Resource Management (on the left menu), select Keys and Endpoint
4. Click Show Keys to reveal KEY 1 and KEY 2.

You can use either of these keys as the API Key. Note that you'll need the key for provider settings

3. Deploy a model

1. Open Azure OpenAI Studio
2. Under Management, select Deployments. Then click Create new deployment
3. Select your desired model under Select a model, and set Deployment name
4. Click Create to complete

Now, your new model should be available. Note that you'll need the Deployment name for provider settings

4. Connect Gasby to your Azure OpenAI

Once you completed the previous steps, you'll have Resource name, API Key, and model deployment name. Then:

1. On the GasbyAI.com site, click on the settings icon (on the top right) to open the Settings popup
2. Under the Settings section, select Azure OpenAI as the Provider
3. Paste your API Key and Resource name.
4. Under MODEL DEPLOYMENT NAME, enter your model deployment created in the above step. Ensure to enter into the right model (GPT-3.5-Turbo or GPT-4)

Under Settings option

1. Choose your chat model
2. If Dall-E is available (you'll need to deploy Dall-E model), check Enable Image Generation*
3. Click Test settings to ensure the settings are correct. Then click Save to complete

Now, you've completed connecting Gasby with your custom API service and should be ready to start using Gasby