H3 Studio 2 — ComfyUI Node Pack + Workflow (v4.0 Hydrogen) by Lora-Daddy
Free For the betterment of my health i prefer this stay fun then be a product. thank you to those who bought it <3 - Donations welcome though! PS. i have yellow fever, i cant stop genning Momo, send help
18/11/2026
multishot and music video overhaul slighty
new max theme for ultimate control. remove windows. (they are hidden at bottom of node when removed) - fade things out. add a background, do all sorts
prompt refinement. + more tweaks custom lighting section, for a more flat realistic raw look. - MAX LAB THEME - LAB editor on the right. > click it to edit anything - 






Description
FAQ
Comments (24)
Did it works with a 16GB GPU?
unconfirmed and personally i wouldn't buy it if i had a 16gb gpu -, cheers bud
Im gonna chime in and say that it... Depends. ComfyUI is very good at offloading into RAM these days. If you have 32gb of ram, keep the resolution low. If you have 64gb, there are not many limitations.
@OhNoMyChickens the hardest thing is the model to run. my node will work on 12b llm but i don't want to promise it will be as a good as the 26b / 31b models
@Lora_Daddy Ah right, the LLM is built-in? Not optional? If it loads, writes a prompt, then unloads then it should be alright. I don't know what you're using, but even a 2-bit quant of Qwen 27B actually does a respectable job with many things.
@OhNoMyChickens ah thats fair, i ts setup to use lm studio, ninfer or google cloud if u have a project id. (trial credits last forever u get like £250, but will reject nsfw often, so give and take.
@Lora_Daddy Thank you for your honest answer.
Is it possible to give this a picture of someone and it analyses the picture and the pose and adds it to the pose bank? Cause thatd be cool
yes thats the entire idea look at the first elsa video - its a custom image thats converted to depth map via the node. <3
lemme upload a small video
Is there an option to pay using BTC? My bank card payment failed.
sorry bro i might try and introduce something like this later at the moment for tax reasons its all going through official channels
@Lora_Daddy maybe a paypal ?
Hello! Sorry to ask, but what is the purpose of this if the node (https://github.com/Lora-Daddy/PromptMasterLD) is down? Does anyone know how to fix this?this an old node i don't use anymore. <3
Gotta say this is my go-to tool now for video generation. It gets better every version. With 2.8 though the sample workflow defaults to a model called 10Eros_Max_h3_8stepTURBO_beta7_test3-native_int8_protected_edges_HQ.safetensors which is gated. Any reason for this?
sorry brother just grab beta 7 from civitai, <3 same thing
Hello hello!
I just got gifted this from a friend and I have to say it is jaw droppingly cool what you have done. I have never used a LM before but I think that I will give it a go now in order to try it out here. Do you have one that you recommend in particular?
pretty much qwen 3.8 or maybe Gemma 4. but its trained towards Qwen3.8-27B-Uncensored-HauhauCS-Aggressive-MTP-GGUF for lm studio.
or using ninfer like my website says if u have a 5090
Bought it, tested it, pretty cool!
The speed is actually good on my mid-end 5060Ti 16 GB w/ 32 GB RAM. Same speed and even slightly faster than most workflows out there.
Where can make request for things such as additional poses?
I'd like to have a "W" variation of "Lies on the back, legs raised in a V" (P055).
Otherwise, is there a way to set this pose and bring variations within the prompt, like "W" instead of "V" and setting the feet in different position/orientation than pointy as in the pose's image?
Thanks!!
in the pose library at the top right you can create or add from an image <3
Oh, that's really nice! Lot of flexibility, love it!
Thanks!
Would buy it if you accepted crypto. Cant link card.
sorry brother, i understand this is just me loosing money , but using itch.io and stripe payments keeps everything official and easy to distribute. appreciate you trying though
