Home
Creators
Search Recent Missing Random
Posts
Search DMs Popular Hash Lookup Tags Random
Importer
Import FAQ
Register Favorites Keys Review DMs Logout
Community
Telegram ThePornDude
Home Creators Posts Import Register Favorites Logout

Happy 421! (Patreon)

Published:
2026-04-21 15:52:48
Imported:
2026-09
Tags:

Content

Hope you were on Cloud 9 yesterday. If you weren't, my 420-friendly futa girls can help you get your fix! Patreon members got 13 new vids and 27 new images yesterday to smoke some bud with.

Free sample because I like you guys:

https://files.catbox.moe/faw80f.mp4

Files

Comments

Dee

Do you know what is the Seedance 2.0 equivalent that creates NSFW content whether it's an image generator or video generator

MomBallsAi

Right now, the best open source models (which is what you have to go to if you want uncensored) for video are Wan 2.2 (what I use) and LTX 2.3. That's about as close as you can get to Seedance 2.0 that is uncensored. The problem for a lot of people is that you either need the hardware (and knowhow to use it) either locally or in the Cloud, or you have to use CivitAI.red to use them.

Dee

1. What cloud are you referring to? Because I might need to buy more Ram and create locally on I guess ComfyUi 🤷🏿‍♂️

MomBallsAi

General "Cloud" there. As in renting a GPU equipped instance from a cloud provider (but then you need to use a template or deploy a Linux/Windows instance manually). I used to use InstaSD, but it is pretty pricey. There is also RunComfy. The nice thing about those two is Comfy is already preloaded, you just have to rent the hardware and install your custom nodes and models and workflow (they provide a lot of this in their templates to run popular models). Vast.ai or Runpod.io are other popular options, but you have to go largely from scratch or use someone else's template. One I considered using was TensorDock, which is nice because you can create long-lived VMs/instances, not ephemeral like the others. I priced them all out back in the day, worked out to about $.50-$1.20 per hour for an RTX 4090 instance at the time depending on the service. In the end, I just shelled out for a new rig with 64 GB of RAM and a 5070 Ti with 16 GB of VRAM, and it can generate a 5 second video in 7-8 minutes. It just ended up being easier than constantly having to screw around with building up/tearing down cloud instances and then I didn't have to worry about being on the clock constantly, I could experiment freely. And it pulls double duty playing video games. Have to find what works for you. There are a lot of options out there. I still keep my 4060 rig (8 GB VRAM) with 32 GB of RAM and it still chugs along generating 5 second videos every 20-30 minutes. So, you actually don't need insane hardware (though you often need optimized workflows and settings that your rig can handle).

Dee

(HP Victus TG02 Prebuilt Gaming Desktop, 12th Gen Intel i5, 6GB Nvidia GTX 1660, 16GB RAM, 1TB SSD, Windows 11) this is my PC right now. So what do you recommend I do next? Because I'm willing to buy more RAM or VRam and download ComfyUi or RunCloud. My main goal is to create ai content for YouTube and try to make extra money on the side while making NSFW content for myself.

MomBallsAi

That is a little on the weak side, but I'd still give it a shot if that's what you have at the moment. If you can upgrade, it'll be faster and more capable. I recommend the best NVIDIA card (newest with the most VRAM you can muster, though I know AMD cards can work, but most AI models are optimized for NVIDIA) you can afford and the most RAM your mobo will support. You'll be a bit limited in terms of RAM and CPU, but the GPU is the biggest factor by far. If you can afford a whole new PC, that's my recommendation. But, if you are just getting started, I'd try to work with what you have or do modest upgrades since it'd be a shame to drop a bunch of money and then decide it isn't for you. Google and LLMs can be great for advice, here is a reddit post of people arguing if you can do Wan 2.2 on 6 GB of VRAM, short answer is probably 'yes' with lots of optimization (lightx2v, few steps, low res, GGUF quant), but it'll still be slow: https://www.reddit.com/r/comfyui/comments/1njvt01/how_do_i_use_wan_22_on_6gb_vram/ . I recommend UmeAiRt's workflows on CivitAI and HuggingFace, this is the one I based my workflow off of: https://civitai.com/models/1824577/img-to-video-simple-workflow-wan22-or-gguf-or-lora-or-upscale-or-fast . GGUF and Lightning/LightX2V LoRAs at 4-6 steps work wonders. Of course, if you are a Teacher Balls member you can just grab my workflows from the archive and use those (as with anything, you'll need to figure out how to install the models and custom nodes). UmeAiRT also provides an installer script which I haven't tried but probably simplifies everything. Good luck!

Dee

Ok so I took in what you said. What are steps to doing ai creation?