While building and experimenting with GPTs, I have found that using GPT Builder / Tester / and the GPT app itself is all under the same usage cap as my GPT. I often run out in the middle of testing and have to restart later but the conversation inside GPT Builder disappears and/or the testing thread has to be restarted. I have been testing my GPTs in the ‘app view’ rather than in testing mode, that way I can go back to various conversations to test updates. There is no benefit to testing mode window where my cap is applied but no memory of conversation is retained. Does anyone have thoughts on this?
Yes, I think if OpenAI opens this for non paid users that would be great or at least to be shared within the organization and the main account should be charged for the total uses.
If your GPT calls an API, debug/testing mode is extremely beneficial, because it shows all the calls to endpoints along with params sent and data received. It’s a lifesaver in debugging GPTs.
Now if only testing wouldn’t count against the cap.
Yeah it is a bit ridiculous that as pro users we are capped to develop by the general use limit. It really is forcing me to do most of my development on a private server with GPT4all. The chat agents are just too limiting in their functionality to struggle with having to stop and return to the project hours later.