There are a lot of puzzle games out there. My vision is a separation into 2 zones - solved and unsolved zone, which you can freely pan and zoom.
Then I realised I don't have the images for it. The solution was to generate them with ChatGPT.
I also generated the main background with it and was shocked how it "sees" the puzzle pieces. No matter what, the LLM did not listen to the requirement or generate the right image. So I decided to leave it with these puzzle pieces. Let people see what a stupid thing it is on the current stage of technology.
I hope you have fun playing!