-
Couldn't load subscription status.
- Fork 113
Description
Hey @jamesrochabrun, thanks for the amazing library and the quick updates! I was just messing around with 4.0.5, and tried to replace my current usage of chat completions in a tool with the new responses API.
I noticed that the official OpenAI client has added a handy output_text property for doing what people do very often for the simplest of requests, deriving the text from a response.
This code from their Quickstart guide shows how they model this property.
from openai import OpenAI
client = OpenAI()
response = client.responses.create(
model="gpt-4o",
input="Write a one-sentence bedtime story about a unicorn."
)
print(response.output_text)I think responseText would be a great computed property to have on ResponseModel, to easily provide that info for tools like mine, which ask ChatGPT for a simple text-based response.
Thanks again, I'm also using your SwiftAnthropic library in the same tool and your work is truly excellent — such a treat from the open source world!
P.S. I didn't want to file a separate issue for this but you have a stray ) in your 4.0.5 release notes.
