Expected Behavior

For the OpenAI client, it is already possible to provide headers for the underlying HTTP Client to set headers for the request. This is especially enterprise setups where you might have an API Gateway or other services in between you and the OpenAI instance in Azure. Those systems in between often require additional or different headers like custom api keys.

For the standard OpenAI client, there exists already a property to configure http headers via config but also via the prompt. It would be great to have the same support for the Azure OpenAI instances.

Comment From: andreas-eberle

Might be a bit related (but as far as I understand, not the same as) #512.