Image Diffusion | Character Consistency | NSFW Content 🔞

Image Diffusion Foundations

From zero to your own ComfyUI image pipeline with Krea 2. Only what counts.

• Your own ComfyUI image pipeline
• A downloadable workflow for every step
• Inpainting, LoRAs and print-ready upscales
• Every setting explained, not memorised

NSFW Pro Pipeline: The PURPLEPINK Workflow, Every Control Measured

Steps, CFG, refine, upscale, four detailers and a 45-LoRA roster, measured on 103 real renders.

• The PURPLEPINK workflow and its 1-click RunPod template
• Every control measured on 103 real renders
• What the four detailers really change, at 100%
• The LoRA stacks the community actually runs

One Face, Every Scene: Train Your Own Character LoRA

Build the dataset, pass the dataset check and train a character LoRA you can use in every image.

• A trained LoRA of your own character
• A clean dataset you can train again
• A way to read the dataset check
• A test to judge and fix every result

One Prompt, Many Images: Wildcards

Write one prompt that turns into hundreds, and stop rewriting prompts by hand.

• One prompt that turns into hundreds
• Layers that always mix well
• Pick, chance and nesting under control
• A wildcard of your own to keep

Seen, Shared, Paid: Prompting and Publishing AI Content

Prompt a persona people believe, publish it on Reddit, DeviantArt and Fanvue, and know the rules, the research and the risks.