Following repeated objections to innocuous wetlook prompts, I ended up with what I believe to be a ban on Microsoft Designer, which is being disguised as "Something's wrong on our end. Please try again later". It's lasted for more than a week now. Something similar happened with Bing awhile back, but at least that was honest about there being a ban.

As I don't know how long the Designer ban will last, and encouraged by soakher's positive comments about Ideogram 2.0, I decided to give it a whirl. A few initial efforts were promising, so I decided to take the plunge and pay for a month's subscription.

One of the issues I've had with every AI image platform I've tried which has produced good quality photorealistic images is that the bottom half of garments can look brilliantly wet whilst the top half stubbornly still looks dry (I'm generalising here). Not with Ideogram 2.0 though, as you'll see here!

If anyone else is thinking of giving Ideogram 2.0 a try, be aware that the free offering gives you few generations per day and that all images are public (and can't be changed to private) unless you have at least the Plus plan, which costs $20US per month. Bizarrely, you also can't delete images unless you're on at least that plan. The good news is that you get 4 images per generation and I've had just 2 images out of all the generations I've done so far (nearly 600) which were blocked as being too risqué. I don't go for scantily clad women though, so it might be a different experience for those who do!

Ideogram also allows words like "clinging" to be used in prompts, compared to the Microsoft engines' almost certain blocking of the entire prompt.

Anyway, I hope the AI wetlook fans out there enjoy this fairly random selection that I've chosen from my limited experimentation with Ideogram so far.

Some of these are really great! I think I need to try this also...

I've also noticed the hair and face often stays dry, no matter how you word the prompt. I may try this and see how it goes. I'm not so much into wetlook, but messy stuff on the face and hair is often ignored by the other AI engines I've been using.

I've been using Ideogram for a while, V1's text imaging and prompt recognition was already very good. Ideogram 2.0 is definitely a big improvement and wetlook looks good with the right prompt. V1 was pretty terrible with most messes. V2 has improved a lot but it still falls short on texture etc.

Thank you to those who've taken the trouble to add likes to the first batch of images. Here's a second selection of images on the same theme. A few of them have a minor muddy element to them for extra realism!

I'm massively enthused by what Ideogram 2.0 has provided so far, so am going to experiment more with images featuring multiple people and/or dresses as the next part of my exploration of what it can do.

can you share some prompts, I tried it and i still can't get them looking nearly as wet as you have in yours.

I've also been using ideogram a lot over the weekend and came to the conclusion that it is able to produce some really great results, way ahead of what Bing is able to do in terms of realism. It has a similiary good understanding of prompts however, making it superior to Midjourney in that aspect.

I've made a few generations over the past couple days, and so far my experience is, that when it's good, it's VERY good, especially for wetlook. I do have a decent amount of failed generations, and it doesn't have as good of a handle on some of the inherent AI issues (hands) as bing, but I'm definitely going to pursue Ideogram further. All these images are straight from the generator, no enhancement or editing done as with many of my images.

Lavette, you seem to have super good ideas about what pictures to make. They are varied and go from pretty to sexy. Great job.

Sorry for the slow reply - I missed seeing your message until now. On other AI platforms, I struggle to get the level of wetness that I'd like on a consistent basis, but have found Ideogram 2.0 to be pretty obliging compared to Bing and Microsoft Designer.

What I do is to make sure that the subject being wet is mentioned in the first clause in the prompt and I tend to use hyperbole (for example "utterly drenched")! I also find that the AI algorithm is more inclined to help out if you give it a reason to want to make the person look very wet. Therefore, for example, you'll commonly see AI generated images for an outdoor location showing it raining or for the area to be flooded. Again from a hyperbole perspective, I won't just say "rain" but "intensely torrential rainstorm"!

As I'm nit-picky about it, I usually specify each of the items of clothing separately, so will add "sopping wet", "saturated", "sodden" or something similar to the desciption of all of the items of clothing. Then I'll round it off by telling the AI to conspicuously highlight the clothes' wetness with a variety of wetness effects, such as trickling water and water droplets.

I know that many people don't go to the same level of detail in their prompts as I do and yet still achieve excellent results, but it's developed into a habit for me to want to be very specific! Of course, if the prompt becomes very long, the AI algorithm gets bored and will start ignoring some of the clauses, particularly those later in the prompt, although some engines actually start to forget what's earlier in the prompt!

Anyway, I hope that gives you some ideas and that you have more success going forward

Ideogram has much less strict filtering of input prompts as well picture results compared to Bing or Midjourney, although you'll sometimes get a cat instead of the Bing dog

It won't give you nudity or NSFW content though, but that's not what I'm looking for anyway. It can give you some pretty sexy results with the right promots.

Additionally, the results are MUCH more realistic than Bing's plasticky and artificial look, at least on par or even surpassing the realism level of Midjourney, with a really good prompt guidance and very realistic wetness effect for almost any style of clothing. I do love casual clothes like jeans, and while Midjourney can sotmeimes do great wetness effects on dresses it usually struggles with jeans.

Overall, I'd say that it's the best AI for creating Wetlook images I have seen so far.