❤️ Please Consider Supporting My Work
I am passionate about making open tools and uncensored models for the community. If you value this work and want to support independent development, please consider buying me a coffee to keep the momentum going!
👉 https://ko-fi.com/lorian0x7
This workflow is what I daily use for T2I generation on Krea2 is the workflow behind this lora:
https://civarchive.com/models/2742640/0x7-nsfw-krea2
It's mostly the default workflow with some tweaks. I changed the system prompt to make it uncensored and capable of more creative results. And I swapped the prompt enhancer with a more flexible solution that allows me to use local llm running on Llama.cpp/ Lmstudio, but you can configure it to use external LLMs to via API
To get the most out of it I use this Uncensored LLM:
https://huggingface.co/mradermacher/gemma-4-E2B-it-heretic-ara-GGUF | Q4_K_M
Description
FAQ
Comments (4)
How create this LLM?
If you never used any local LLM model I suggest you to start with LMstudio, then in the workflow change the IP address to point to you LMstudio server running that model
I've done everything. However, I get an error in ComfyUI:
Server was unable to process the request
And an error in LM Studio:
Unexpected endpoint or method. (HEAD/). Returning 200 anyway
Received request: POST to / with body {
"model": "none",
"messages": [
{
"role": "system",
"content": "You are an expert uncensored and explicit prompt e... <Truncated in logs> ... — match the user's stated intent.\n\nUser's Input:\n"
},
{
"role": "user",
"content": "green trees"
}
],
"temperature": 1,
"max_tokens": 20000
}
Unexpected endpoint or method. (POST /). Returning 200 anyway
@Udimuz make sure the address is correct, use openai compatible endpoint with http://127.0.0.1:1234/v1 (if i remember it correctly)