Prompting images in ChatGPT 4

I’ve been trying to render sequential images in GPT 4 but haven’t been able to successfully do so without errors.

What I mean is suppose you want to render 3 images depicting a story of a man who drives to a gas station but realises when he gets there that he forgot his credit card at home. What you usually get are 3 different looking men, wearing different clothes and driving three different looking cars.

Is there a way to render images so that the same character and key objects appear the same in all three images. If not, how can you minimise discrepancies in images that are supposed to be used in a sequence?

1 Like

There’s already a Topic covering this:


I have checked this out and it’s not really clear what the contributors are doing to get consistent images. Could someone please explain their process clearly in numbered steps to take?