Thanks for the feedback. We wanted to make the interface as simple as possible to use but offer advanced settings to really fine-tune your image prompts and get the outputs that you want. Understand that some of the settings are difficult and we're working on a whole new video series to make this more simple. I totally get that building out the prompts can be difficult compared to other tools, we're also working on an AI prompt assistant at the moment and have tutorials on the image 2 prompt tool coming which will showcase how you can get certain images quicker.

In terms of speed and reliability, we keep trying to get our reliability better. In the past it hasn't been as good as we wanted, some of the recent improvements have now seen our % failure rate drop to an all time low of 0.06% and we're actively looking at those small fails still to reduce further. Speed is something we're always looking to improve. One thing with Supermachine is that we don't limit the amount of images in the queue so a user can queue up multiple generations at a time. We see this as a strong positive feature but it can mean if a user has queued up multiple it can make the queue longer. We're looking at ways to tackle this in future but the queue rarely gets above 3 generations long.

I think a lot of the success with a tool like this is the learning of making the solid prompts and there is definitely a way to get awesome images with our technology, it is the same technology that other tools are using so I think with the AI prompt improver and better documentation/tutorials we can showcase the best way to use Supermachine and have success.