We need to have an easier onboarding experience for the openai-gpt agent to ensure there is at least one agent that users can use with little barrier of entry. To do so, we need to prompt users to choose between Azure OpenAI or OpenAI as their endpoint of choice. Then prompt them for any required values for each option.
User flow
PS>aish
(user choose openai-gpt agent)
Which openai service would you like to use?
> Azure OpenAi
openai
(user selects Azure OpenAI)
Please input the endpoint URL of your Azure OpenAI Deployment, for more information on how to create an Azure OpenAI Deployment please refer to aka.ms/openai-gpt-aoi: <My Endpoint>
Please input the name of your deployment: <Deployment name>
Please input the model name: <model name>
Please input your API key: <key>
aish:1>
For OpenAI:
PS>aish
(user choose openai-gpt agent)
Which openai service would you like to use?
Azure OpenAi
> openai
(user selects OpenAI)
Please input your API key: <my openAI API key>
aish:1>
@StevenBucher98 and I discussed about this issue, and agreed that it's better to lead the user to the document about how to configure GPT instances for the first-time use.
Summary of the new feature / enhancement
We need to have an easier onboarding experience for the openai-gpt agent to ensure there is at least one agent that users can use with little barrier of entry. To do so, we need to prompt users to choose between Azure OpenAI or OpenAI as their endpoint of choice. Then prompt them for any required values for each option.
User flow
For OpenAI: