Hi ! What's the difference between the AI gererated text card and my own GPT api please ? Imagine I use my chatgpt API with gpt-4 model, and I use the same model with the 'AI generated card'. All have the same amount of temperature, tokens, same prompt/instructions, etc. Is my own chat gpt API still provide better answer in general ?

I wouldn't say "better" responses per se... The chat completions API provides many lower level options you just don't have access to in botpress, such as tool calling, stop sequences, json mode, etc If you're just using it for simple text generation it's practically the same

Hey @magnificent-byte-37082 , the short answer, there is no difference. Long answer is: We add wrappers from our side around the AI cards, to be able to control it's inputs and outputs, do parsing if needed, etc. So the output from the OpenAI might be different from the output you get from our AI Generate Text card.