I've been trying out Dreamup AI through deviantart. While I've made some fat creations, does anybody have any tips on what terms I could type in to make them look closer to the 400-500 USSBBW range?

Any tutorials anywhere on adding speech bubbles to pictures in general and recommended fonts? I can always prompt for a speech bubble but then i'm at the mercy of how big the AI decides to make it. I normally use GIMP for touching gens up after words but I dont see a module or template anywhere in there for speech bubbles unless i'm just really blind.

>>18540 I’ve just started dipping my toe into this stuff so take it fwiw, but I’ve had limited luck combining terms that mean “really fat” — “extremely fat heavyset morbidly obese” etc. That still gives random sized results. One prompt will give me a 300 lb woman, the next a plausible 450 pounder.

What prompt would I have to use if I wanted to make a woman/celeb look like she were swelling up with fat or inflating like a balloon or fully round and ripped/ripping out of her clothes? And what if I wanted to apply inflation to specific body parts? Also, what AI would be best for that kind of thing?

Faceapp did away with the "face swap" feature. Is there an app I can download that can replace that?

>>19089 Try Facetool-AI in the app store. It gives you 24 free face swaps per day and does a pretty good job. It's a bit buggy sometimes though.

I'm making a video2video workflow for comfyui to do video morphs, how do I get rid of the flicker? Here's the workflow https://pastebin.com/EzVwzast

Anyone know any free "nudify" sites that work for fat girls?

Hey, so I have about 60GB of fat arts from heaps of people on twitter, dA and pixiv that I've been scraping through 2023, for the purposes of making LORAs. However, I have no idea where to start, what settings to put, and am too retarded to really do this stuff without a step by step guide. Could anyone please point me in the right direction? an example of one of my datasets: 150 pngs of varying dimensions in one artist's style, curated to only include blonde fat girls. How many steps? tf do I do with the tagging? Looking at the training tab in the SD WebUI giving me an aneurysm.

>>7250 >>7251 kinda funny how using inpaint i have the exact opposite issue, i prefer bellies with flab and rolls and it mostly gives me big bloated ones

QUESTIONS: -is ai capable to make inflation art? -which site allows you to make your own fat art?

I had a pretty good run using bing for a while but recently the content filters were updated and my old prompts don't work any more. Even the stuff I could get previously was severely limited. I have tried stuffer.ai and it's pretty bad. Light-years behind bing. I also set up a dalle3 account and tried generating from the CL but I've found the quality to be much worse than bing, despite both running the same backend. What in you opinion is my next best option for AI generated SSBBW? TIA

hello anons, I have been fucking around with stable diffusion, i managed to get something like this after a while. im going for realtively realistic immobility or near immobility. you will probably tell me watch some tutorials but i'd rather just ask what you'd do as a next step and how you'd go about it. say i wanted her belly to rest on the bed, i tried with the inpainting but that just kinda adjusts the edges but doesn't actually change anything big. I wanna keep the belly button and make it look like there's cum dripping out which in this case was just a lucky gen. have her arm and hand rest on her belly fold i also couldn't manage, it just removes her arm entirely. do i hop into photoshop and change around things there and then have the ai fix it? this is a bad case of just not wanting to read i must admit. but if you like this i want to make more of it so help me out and teach me the craft please.

>>20666 im slowly figuring this out we're making progress. need to start the details soon now,

>>20666 >>20729 I wrote a guide on my process which you can find here. It's not even a 20 minute read, and you can follow along as you're working on an image: www.deviantart.com/theguywhodidathing/art/How-did-you-DO-that-989204336 The broad steps are as follows: Generate an image at a low resolution. Edit that image to get closer to what you want to see, then reprocess it in img2img multiple times, slowly increasing the resolution. Edit the resulting images to help guide the process, adding and removing LoRAs along the way. Touch up the final image by reprocessing specific parts like hands and faces, then edit the final result like you would a photograph. In your specific case, here's what I did: The first image is just what happened after I ran your image through a blend of checkpoints (more info in the guide). Second image is what happened after I did my 'process'. I scaled the image up in PS a few times by 1.2x, each time reprocessing the image 4 times (at around .4-.48 denoising strength) and picking the best looking option. I only used the Symix LoRA at .5 strength the entire time. Once I had it around 50% of target resolution, I scaled the image up by 2x in Photoshop, then selected areas that needed attention: face, hand, feet, and belly. When I reprocessed the belly, I painted a couple of grayish/yellowish marks around the navel and used the Cum On Any LoRA (https://civitai.com/models/24269) including the words Cum and Cum in Navel. I layered on each of the reprocessed parts, masked out the edges, and that's the second image. This final image is after tweaking exposure, contrast, saturation, and skin tone. That's just photoediting though, and definitely not needed. Hope this helps.

Ok so how can I turn my wife into a blueberry? Obviously using her pics. Basically put her in wonkas factory, or infront of an eating machine, like a kinda deep fake ai type thing if that makes sense? Thanks in advance

Hi, wanned to start doing ai art and went to the guides thread, but some of the links there are outtaded or expired. Does anyone here has a up to date 2024 installation guide or video on how to do bbw art? Thanks

Hey all. Tried looking through this thread and searching through the catalog to no avail; not sure where else to look for a comprehensive explanation. What is the best chatbot available currently, given the criteria of "good memory", "ease of character creation/customizability", and "not NSFW filtered"? I have used C.ai since late 2022 but have become frustrated with things I can't accomplish due to either censoring, or me being able to draw things out *without* censoring and have them go along with it, only for them to forget the scenario we're currently in. In a discussion with a friend last year, they told me Venus.ai let you do neat things that were explicit, but that its memory and response quality did not match c.ai. Basically, I want NSFW chatbots, but not at the cost of overall decreased quality of conversations or interaction, and I am willing to pay for this. What are my options?

Can somebody make a new gen furry thread im too retarded

does anyone else get a weird grayish hue to their images? It always seems to happen to me at the end stage of the creation process.

I've been thinking about a software that could be downloaded to import ai character jsons/cards into a software for some weight gain ai bots since I didn't really want to deal with the restrictions and paywalled stuff for some chatbot websites like Character.ai or CrushOnAI for example. If there is any softwares I could use that would be nice.

Anyone have a guide to making expressive faces? Or less that they use with SD 1.5, or prompts in PonyXL? ScarebroArt pumps out pretty high quality pictures with fantastic facial expressions, and I'd love to be able to get anything close to these

What is the best model for immobile size so far?

Is there a way to prompt SDXL or more specifically PonyRealism to just make the belly bigger? Because it makes the rest of the body fatter too. Ok, i might have just answered it myself, the pregnancy tag. And I used it too at times...I need more sleep.

>>25542 Round belly, stuffed belly, big belly, bloated belly all will typically make the belly bigger without adding fat elsewhere. With symix anyways. If you need to prompt pregnant for just a fat belly, make sure you add Linea nigra to the negatives to make it not look overtly pregnant.

Can anyone recognize what software/models this guy uses? https://www.deviantart.com/xxeatmexx/art/The-girl-from-the-supermarket-SSBBW-3-1071438459

can yall people who know what you're doing do something with this? its all about that perspective eventhough i fucked it already. how do i get correct perspective to work? also generally what would be the next steps here to make something outta this?

Is there another ai program that allows you to generative fill a selection and use a prompt without any hard restriction? I used to use problembo but their new model sucks now

does ai understand prompts in regards to the "camera" we are viewing the image or video from? Like does it understand "walking towards camera" or "facing the camera"

Has anyone had success prompting for stuckage? I've tried a bunch of different approached (most recently in PonyRealism) but can never get it to work right. I don't see much art of it either, so maybe there's no loras for it either. If anyone has had success then any pointers would be appreciated. Here's some failed pictures that got close.

>>27108 Your "fails" are better than my best ever.

>>27127 PonyRealism is very forgiving, at least in my experience. Maybe it's just because I've messed around in Realism more, but I find it easier to prompt than PonyXL, at least in terms of getting what I want

>>2561 Is there any AI/website that can do a reverse-prompt-generation for a given picture? i.e. giving artwork or photos as input and returning as many descriptors as possible in the form of a tag (bonus points if it includes some sort of percentage of likelihood). i.e. giving a picture like this and expecting a list of tags in return (woman, bent_knees, midriff etc.)

>>27185 Yeah, I think AUTOMATIC11111 comes with a built-in Danbooru/CLIP reverser. Don't expect it to work miracles but it can give you a headstart

>>27108 try training a lora, maybe on civitai? you can also do img2img to gen some content if you don't have much source material.

Can anyone tell me what checkpoint may have been used to generate this? It came from this Perchance generator https://perchance.org/lunatics-full-extras-no-limit-ai-the-have-fun-gen-full-with-extras Which is a great generator btw, I use it to make a raw image then edit in SD but I'd like to have whatever checkpoint or merged checkpoint was used to make it. Thanks for any help!

>>23949 Try to lower your CFG value

I’ve had the account for ages, and I want to get rid of it since the filter is absolutely terrible. It’s ruined the experience for me. But what I want help with is how to I get rid of the account that’s linked to my google shit? I was stupid, and had no idea what I was doing, and I selected “Continue with google.” Now it seems like everything time I “delete” my account, and test if it’s still around by logging into character.ai, the account STILL exists, even after months of trying to pretend it doesn’t exist. Help? Please? Would mean a lot :(.

OK, I just tried to make my first animations using Kling last night. At this point they've been "creating" for over 16 hours. Is it stuck or does it really take that long?

>>28217 Update: after waiting over 48 hours, I gave up and exited Kling. That sucked.

Can someone post full sequence pls?

>>2561 First, a tip: I've been getting great results with stuffer.ai using a reference image of the non-trained character I want to get, and this site: https://huggingface.co/spaces/hysts/DeepDanbooru Which guesses the most appropriate characteristics of a picture in danbooru tags. Picrel is Alena, getting bloated using the settings of some other OC. Secondly, I want to ask the following: 1) What kind of tips do you guys have to get an even more precise rendition of a character/pose you want to get? Sometimes it feels the AI just ignores everything I tell it to unless I specifically copy settings off of someone else. 2) Is there some sort of list of all the available artists/styles I can ask the AI to copy, or at least the most consistent ones?

Guys, i usually use Automatic111, and i heard what this is not to great, compared to ComfyUI, how can i make transision to this system? I need to reistall everything, or i can change something in A111 folder?

>>28433 Comfyui doesnt use gradio like A1111 forks do and uses its own source code so you'll need to install comfy to its own folder, it doesn't take up much more space than a1111. There is a file in the comfyui directory that you can tell it to use the lora/checkpoints folder in your a1111 directory if you want to use both side by side until you decide which one you like best. You may want to look into Forge instead maybe. I like comfyui because it gives me a very autistic level of sliders and dials, but thats not for everyone and might end up turning you away if you dont figure out how to setup your own workflows. Forge has some nice QoL updates and is structured similarly to A1111 as its a fork.

>>28480 Thx for info <3

What's the best storytelling AI without filters? Also, what's the best one to try to circumvent thr filters on? I noticed my favorite author's stories were written with help from AI and he's barely released anything in the last couple of years and I want to see where the stories go I really liked AI Dungeon when it was uncensored and unlobotomized.

I'm trying to find at ai that was made of the character from oopic-oorts don't be picky comic but can't find the link anymore. Anyone have it still?

A couple months ago, there was a pony based on Herohero Tom's style. It's since been deleted from Civitai. Can anyone share it?