Skip to main content
terribleweaver
terribleweaver
Participant
February 25, 2026
Question

Same image is begin made over and over again

  • February 25, 2026
  • 1 reply
  • 38 views

On many image creation libraries like Google and Adobe, no matter the prompt it produces the same image over and over again with no changes but it charges credits for it. It’s absolutely terrible.

For instance I make an image of a house then I go to change some stuff on it and prompt it but it doesn’t listen to the prompt at all it give me the same same image and I lose credits wasted time and energy and credits. 

At least there should be a fail switch and returning credits because this is not acceptable by any means.

    1 reply

    Kartika Rawat
    Community Manager
    Kartika RawatCommunity Manager
    Community Manager
    February 26, 2026

    Hi terribleweaver!

    I’m sorry about the issue. Could you share the prompt and the generated results to see what’s happening? 

    Let us know. Happy to help.

    Thanks,

    Kartika

     

    moved to questions

    RSUMG
    RSUMG
    Participating Frequently
    March 30, 2026

    Here’s two examples where there are significant differences in the prompt and the image returned is almost exactly the same. It seems to be stuck on this output now :(

    https://firefly.adobe.com/open/urn:aaid:sc:US:47737ead-8c8f-49d8-8750-0eb51b306ea6

    https://firefly.adobe.com/open/urn:aaid:sc:US:7cdbf113-fc1e-4175-a1ae-a19ad14bbd27

    Kevin Stohlmeyer
    Community Expert
    Kevin StohlmeyerCommunity Expert
    Community Expert
    March 30, 2026

    Hi ​@RSUMG - the examples given have both a reference image and the same base prompt used in both examples. The prompt appears to be GPT-driven.

    Adding “no text, no people, no objects, no photography, no recognizable natural scenery, no clouds, no mountains, no grass, no ocean, no sun, no harsh contrast, no chaotic texture,” does nothing for Firefly - it ignores all “No” values so those additions are null. 

    If I remove those, you have the same base prompt as the first example. 

    With both inputs left as is - you are seeing variations, but they are subtle because you are too specific with the instructions. You can see the colors were accounted for and did give you the requested changes. 

    If you edit the prompt manually you can get more variations in the results or remove the reference image to get rid of structure. 

     

    Terms & ConditionsAccessibility statement
    Using the community Terms of use Privacy Cookie preferences Do not sell or share my personal information ad choicesAdChoices