I know it is a little late, as it is over a week ago... but it does depend on which one you went for. If you want free and decent, there are a couple that I have used:
https://mage.space this has good free functionality, it is fast, accurate, with lots of momdels to choose from. I personally subscribe to this, as it allows uploads of LoRAs (fine tuning/additional 'knowledge')
https://sexy.ai good free generator, started with this one... models are limited, but is capable of good images
All images created with the same prompt: beautiful goth woman, long thick hair, HDR, high resolution, (detailed face), ((sheer robe, partially open)), (crawling through deep mud on hands and knees)
and same settings: Steps: 30 Strength/CFG Scale: 7.4 Negative prompt: Asian (deformed, distorted, disfigured:1.3), poorly drawn, bad anatomy, wrong anatomy, extra limb, missing limb, floating limbs, (mutated hands and fingers:1.4), disconnected limbs, mutation, mutated, ugly, disgusting, blurry, amputation
I couldn't use the same model, as the free accounts on each offer different models Sexyai: realistic vision Mage.Space: SDXL
When playing with models, some things to bare in mind: 1. sometimes less is more... reduce steps and cfg strength/guidance as well as increase... see what impact they have. I have had some amazing results with guidance at 0.5 and strength all the way to max.... it is very much model-related annd how it was trained. 2. orientation of the image seems to matter. if you have a landscape image a majority of the generated images will be laying down... portrait standing up. 3. use bracketing and punctuation... play with it... sometimes it helps to bracket things that could be interpreted as separate. so: "woman, crawling in mud" could be different from "woman, crawling in mud", which could be different from "woman (crawling in mud)".
Hope that some of this helps... and good luck, look forward to seeing some results if you try them.
First two images are mage.space, second two are sexy.ai
Love you, too
