Options
Go to last post Go to first unread
ADP001  
#1 Posted : Tuesday, August 11, 2026 7:25:55 AM(UTC)
ADP001

Rank: Advanced Member

Joined: 8/2/2016(UTC)
Posts: 72

Thanks: 1 times
Was thanked: 14 time(s) in 13 post(s)

Even though Z-Image and Flux handle regular text just fine, it’s still worth knowing which specific keywords are available to achieve specific effects. You can simply incorporate them into the text or list them at the end as a keyword list.


Sometimes one simple keyword can make the difference.


Here’s a link to one of the many online cheat sheets:



https://sheetly.org/chea...stable-diffusion-prompts


PedersenDraw  
#2 Posted : Wednesday, August 12, 2026 9:05:29 AM(UTC)
PedersenDraw

Rank: Member

Joined: 12/14/2025(UTC)
Posts: 18

Thanks: 31 times
Was thanked: 8 time(s) in 7 post(s)
Thanks a lot for the helpful tip!
topspin  
#3 Posted : Tuesday, September 1, 2026 4:26:06 PM(UTC)
topspin

Rank: Advanced Member

Joined: 12/7/2015(UTC)
Posts: 98

Thanks: 80 times
Was thanked: 58 time(s) in 45 post(s)

Originally Posted by: ADP001 Go to Quoted Post


Even though Z-Image and Flux handle regular text just fine, it’s still worth knowing which specific keywords are available to achieve specific effects. You can simply incorporate 



More complex than that, particularly for erotica.


Flux -- the base model, released by Black Forest Labs -- doesn't do X rated, doesn't respond to a prompt like "blow job"


To get that, if you want to do it through prompting, you need a fine tuned model (or a LORA) which has that particular "skill" and associated keyword.

So with erotica like from the getPhat series (models for Pony, SDXL, Flux, and now I think Krea); that's got its own list of X rated keywords; they are model specific. 

That's why you'll find that prompts mentioning spanking may not do much of anything . . . if its not trained on the concept for that particular _version_ of the model, the "keyword" doesn't mean much. You can still get spanking images by using other techniques, but if the keyword isn't associated with a visual concept in that particular model, thn it doesnt do anything


As a strained analogy -- in some of the DazPoser models, there are no genitals. You have to have appropriate geometry -- either in a built in model, or in a geograft or similar -- if you want hardcore. AI tools are analogous, if you want X rated output you need somemething that's trained on X rated concepts (though there are "tricks" that you can sometimes use to get erotica out of, say, Google Gemini or ChatGPT). 
 

Edited by user Tuesday, September 1, 2026 4:29:28 PM(UTC)  | Reason: Not specified

ADP001  
#4 Posted : Tuesday, September 1, 2026 6:12:46 PM(UTC)
ADP001

Rank: Advanced Member

Joined: 8/2/2016(UTC)
Posts: 72

Thanks: 1 times
Was thanked: 14 time(s) in 13 post(s)

Time has changed. Today there are lots of Flux models with "NSFW" onboard. No need for LoRAs (but you can; there are a lot out there, even more specialized ones; but a lot of the old ones may not or not properly work with the new models).
In addition, these models use an "uncensored" LLM to translate the prompt. In other words, it understands prose (long, rambling texts). And the better the prose, the better the result :)


https://civitai.com/models/2308682/qwen-3-8b-flux-klein-9b-te
https://civitai.com/mode...m?modelVersionId=2474529

You may find all relevant models on civitai (not only Flux). But you must register (free) to see NSFW (uncensored) material.

And, as I described in some other posts, Flux is easy to use with "Stability Matrix" or "InvokeAI" (after the basic installation: just one click after writing the prompt)

Edited by user Tuesday, September 1, 2026 6:15:46 PM(UTC)  | Reason: Not specified

dsv4600  
#5 Posted : Wednesday, September 2, 2026 3:02:07 AM(UTC)
dsv4600

Rank: Advanced Member

Joined: 12/18/2012(UTC)
Posts: 123

Thanks: 63 times
Was thanked: 40 time(s) in 35 post(s)

Originally Posted by: ADP001 Go to Quoted Post


Time has changed. Today there are lots of Flux models with "NSFW" onboard. No need for LoRAs (but you can; there are a lot out there, even more specialized ones; but a lot of the old ones may not or not properly work with the new models).
In addition, these models use an "uncensored" LLM to translate the prompt. In other words, it understands prose (long, rambling texts). And the better the prose, the better the result :)


https://civitai.com/models/2308682/qwen-3-8b-flux-klein-9b-te
https://civitai.com/mode...m?modelVersionId=2474529

You may find all relevant models on civitai (not only Flux). But you must register (free) to see NSFW (uncensored) material.

And, as I described in some other posts, Flux is easy to use with "Stability Matrix" or "InvokeAI" (after the basic installation: just one click after writing the prompt)



No, topspin is absolutely correct. Even with the so-called "uncensored" models you will find it one hell of a job to generate decent erotica without the use of LORAs in the chain. Z-Image, in particular, has a dreadful understanding of human genitalia and is very good at having a penis floating in mid air or attached to a gender that you didn't ask it to be attached to. And that doesn't even touch on the whole issue of actor positioning which is a minefield. I actually gave up with Z-Image for just that reason. CyberXL and Pony do a much better job. In fact, having a locally installed AI is good if for no other reason than to learn and work through all those issues, without flushing expensive generation credits down the toilet with an online site - if you can even find one these days that isn't "Morality Locked" or requires your identity documents to be uploaded and shared with every scammer on the planet.

What really amuses me about AIs though, is that on YouTube now they are advertising for people to learn how to use AIs like Claude, but the whole point of an AI is that it does things for you that are too difficult or time consuming for you to do yourself. If you have to learn how to interact with an Artificial "Intelligence" then it is not an Intelligence at all. It's a programming language.

Welcome to the future.

ADP001  
#6 Posted : Thursday, September 3, 2026 4:24:37 AM(UTC)
ADP001

Rank: Advanced Member

Joined: 8/2/2016(UTC)
Posts: 72

Thanks: 1 times
Was thanked: 14 time(s) in 13 post(s)

@dsv4600


All of my current images were created using Z-Image or Flux. Even if, in your firm opinion, that simply can’t be the case.


https://www.renderotica.com/artists/adp001/Gallery



I have a little over a terabyte of models (Z-Image, Flux) stored on an SSD here. Plus about a hundred LoRAs on top of that. Some of these models definitely don’t excel in the NSFW category. But they’re much better at other things.
And yes, some of them work better with one LoRA or another. Especially when it comes to bondage and the like.

The claim was that Flux or Z-Image don't understand the prompt if it contains NSFW content.
That may be true for some online services.
With a local installation, however, this isn’t the case if the prompt is labeled NSFW or “uncensored.” A customized mini-LLM ensures this by analyzing the prompt (and enabling prose generation).

The fact that the quality of the depicted genitals doesn't always live up to expectations: My God, that's what LoRAs are for! That's what they're there for!
It’s just wrong to say that LoRAs would need to come with the appropriate “keywords” to even understand an NSFW prompt in Flux or Z-Image! The “text translator” takes care of that (it works very differently from SDX, Pony, etc.).

One thing that’s 100% certain is that most SDX or Pony prompts won’t work out of the box with Z-Image or Flux. Even the way you control the camera varies.

ChatGPT or Perplexity can help clarify the differences.

If you use a local LLM, you need to keep in mind that most LLMs are unaware that there is an uncensored version of Flux and that Z-Image is, in principle, NSFW.
The training date for locally used LLMs is usually earlier than the release date of free NSFW models. What you can do, however, is train the local AI to be smarter. Then it will be able to generate top-notch NSFW prompts as well. For example, it will also be able to handle a request like, “I’d prefer to have that person in the other corner,” using completely natural language.

And if you use a larger LLM model locally, it can also load and analyze the generated image, thereby generating better prompts (smaller LLMs can do this too, but they don’t analyze it closely enough to capture details—they simply lack the memory for that).
It goes without saying that the large online LLMs (such as ChatGPT and the like) don’t handle NSFW content.

Edited by user Thursday, September 3, 2026 4:53:38 AM(UTC)  | Reason: Not specified

Users browsing this topic
Guest
Forum Jump  
You cannot post new topics in this forum.
You cannot reply to topics in this forum.
You cannot delete your posts in this forum.
You cannot edit your posts in this forum.
You cannot create polls in this forum.
You cannot vote in polls in this forum.

Powered by YAF.NET | YAF.NET © 2003-2026, Yet Another Forum.NET
This page was generated in 0.094 seconds.

Notification

Icon
Error