Instructions For GPT HTTPs
Uploaded by
fihinih321Instructions For GPT HTTPs
Uploaded by
fihinih321OpenAI is updating and evolving rapidly. Sometimes the updates come faster than Make’s.
As
such, it can be necessary, at least for testing, to make raw API calls while waiting on Make to
update their modules.
This has happened a couple of times now with the upgrade from GPT-3 to GPT-3.5 and even
GPT-4.
To prevent you from having to wait on Make updates in the future, I’ve created this scenario that
simply makes HTTP calls to the OpenAI API. Currently the “model” is set to “gpt-4” but that can
be changed.
Simply input your API key and edit the tokens and temperature you want to use.
This scenario also only includes two “back and forth” chats. One prompt input, then an output,
then another input.
To add more, simply add an additional ,{“role”: “user”, “content”: JSON string} to the JSON.
Copy the modules and paste them into any other scenario to use this feature.
For more information, also look at OpenAI’s API documentation here:
https://platform.openai.com/docs/guides/chat
https://platform.openai.com/docs/api-reference/chat
