Hi Rirriam,Thanks for your feedback, it helped me uncover a transient bug. Sometimes the wrong model was used meaning that you would get images in a completely different « style » for the same prompt. It is now fixed. I’m sorry that generation time is long when a lot of users are using the app at the same time. The only way to avoid that would be to scale up the infrastructure but that would increase the costs significantly, I have to made the tradeoff to have longer generation times during peaks but keep the price reasonable for a tool producing images with this level of quality. I hope that you understand that choice.Best,Thomas

Hi Kurdsboy, 1. Being able to keep a constant low generation time would translate into a significantly higher price for the users to offset the additional GPU costs. I'm thinking of creating a higher paid plan for users that want to get under 10 seconds generations even when traffic is high2. That is in the roadmap, thanks for the suggestion3. Same as 2 :) I hope you'll be pleased with the next version

Hi Red hair89,I’m sorry that generation time is long when a lot of users are using the app at the same time. The only way to avoid that would be to scale up the infrastructure but that would increase the costs significantly, I have to made the tradeoff to have higher generation times during peaks but keep the price reasonable for an tool producing image with this level of quality. I hope that you understand that choice.Best,Thomas