Azure OpenAI Deployments Support

[mputerity]

Can you add support for custom Azure Open AI deployments?

Just setting the env vars like:

OPENAI_API_TYPE=azure
OPENAI_API_VERSION=2024-09-01-preview
AZURE_OPENAI_API_KEY=...
AZURE_OPENAI_API_BASE=...

and using

agent = Agent(
model=<my_deployment_name>,
name="Assistant",
instructions="You are a helpful assistant.",
)

does not work.

[alexprengere]

I believe this has already been answered in #44
I agree that this could be more prominent in the docs. What worked for me was the solution in this comment. Perhaps there is a way to use directly the responses API, to avoid having to pass model=OpenAIChatCompletionsModel(...) for each agent.

[alexprengere]

For the record the openai Python library does not currently support Azure OpenAI deployments for the responses API, so it looks like model=OpenAIChatCompletionsModel(...) is mandatory for now.