Download
1 variant available
Join The Tinkerer on Whop. Membership gets you early releases, private tools and a bunch of extra stuff.
👉 Join on Whop
💬 Join the community for support, free tools and early news on Discord
V1
A clean and practical ComfyUI workflow built for CyberRealistic Krea 2.
This workflow is designed as an easy starting point for creating high-quality stylized images.
Workflow files
This upload includes both the JSON workflow and a PNG workflow image.
Use the PNG version if you want the workflow to load with the original node layout, group positions, and visual formatting preserved.
Custom Nodes Used
Everything else in the workflow is stock ComfyUI. These nine packs are required:
rgthree-comfy — Context Big, Seed, Power Lora Loader, Image Comparer, Fast Groups Bypasser, Label
Handles the wiring and UI of the workflow: context bundles, the LoRA stack, the before/after comparers and the module toggles.
https://github.com/rgthree/rgthree-comfyRES4LYF — ClownsharKSampler_Beta
Advanced sampler suite with its own solvers and schedulers. Used for both the 1st and 2nd pass.
https://github.com/ClownsharkBatwing/RES4LYFcomfyui-negpip-zimage — ZImageNegPipPrompt
NegPiP-style negative prompt weighting. Merges the positive and negative strings and converts the negative side into negative weights before sampling.
https://github.com/cyberdeliaAI/comfyui-negpip-zimageComfyUI-Krea2T-Enhancer — ComfyUI-Krea2T-Enhancer
Patches the Krea 2 text-fusion path during sampling to improve prompt adherence. One MODEL in, one MODEL out. Still marked experimental by its author.
https://github.com/capitan01R/ComfyUI-Krea2T-EnhancerComfyUI-SeedVR2_VideoUpscaler — SeedVR2LoadDiTModel, SeedVR2LoadVAEModel, SeedVR2VideoUpscaler
Diffusion-based upscaler for video and stills. Models download automatically to ComfyUI/models/SEEDVR2 on first use. Optional, see the note below.
https://github.com/numz/ComfyUI-SeedVR2_VideoUpscalercomfyui-vrgamedevgirl — FastUnsharpSharpen
Fast unsharp masking, used to restore edge detail after the upscale. Part of a larger video enhancement pack.
https://github.com/vrgamegirl19/comfyui-vrgamedevgirlComfyUI-Impact-Pack — FaceDetailer, SAMLoader
Detection and detailing toolkit. FaceDetailer crops the face, resamples it and composites it back into the image.
https://github.com/ltdrdata/ComfyUI-Impact-PackComfyUI-Impact-Subpack — UltralyticsDetectorProvider
Separate install that provides the YOLO detectors needed by the FaceDetailer bbox input.
https://github.com/ltdrdata/ComfyUI-Impact-Subpackcomfyui-cyberdelia-metadata — SaveImageWithMetaData
Drop-in replacement for Save Image that writes clean A1111/Forge-compatible metadata into the PNG, with Civitai compatibility fixes.
https://github.com/cyberdeliaAI/comfyui-cyberdelia-metadata
Notes
Impact Pack and Impact Subpack are two separate installs. Installing only the Pack will leave the face detector missing.
The Upscale, 2nd Pass and Face Detailer groups are bypassed by default. The workflow runs out of the box without the SeedVR2 or SAM/YOLO models on disk. Enable the modules you want with the toggle node in the Models group.