Nynxz's Krea 2 Edit (NK2E)
An in-context image-editing LoRA for Krea 2. Give it a source image and a text instruction (like "give her pink hair" or "add sunglasses") and it edits the image.
Experimental and unofficial. Results are rough and inconsistent, and vary a lot by edit type. Not affiliated with or endorsed by Krea.
Links
- ComfyUI Node Pack: https://github.com/Nynxz/ComfyUI-NK2E
- Training Toolkit: https://github.com/Nynxz/NK2E
Use (ComfyUI)
- Download a LoRA from
comfy/v0.1/into your ComfyUImodels/loras/. - Install the node pack (link above) and load its example workflow.
Strength around 0.7 is a good start.
Versions
| Version | Date | File | Rank | Training data | Notes |
|---|---|---|---|---|---|
| v0.1 | 2026-06-30 | comfy/v0.1/NK2E-v0.1.safetensors | 64 | ~37k edit pairs (MagicBrush, OmniEdit, InstructPix2Pix) | recommended |
See CHANGELOG.md for history.
What it does well (and not)
- Good at: keeping identity and composition, localized edits (hair, accessories, objects), and global style or lighting changes.
- Weaker at: recolouring a whole large region, and scene or structure changes (swapping a background, adding a person, re-posing).
It is not a mask or inpaint model; identity is preserved without a mask.
Training a specialist on top
comfy/ holds the ComfyUI-format weights for inference. train/ holds the native-format
weights for training. To train a narrow specialist on top of this generalist, point the
toolkit's base_lora at train/v0.1/NK2E-v0.1-train.safetensors. See the training
toolkit (link above).
License
Krea 2 Community License (see LICENSE and NOTICE); commercial use only under $1M/yr revenue. Not an official Krea project.