# Best way to combine pose, identity, clothing and background in Krea 2 ComfyUI? — retained record

> Retained reddit record from reddit:StableDiffusion.

Canonical: https://highsignal.app/data/records/7aeac261adac1975

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:StableDiffusion)  published 2026-11-03 06:48 ingested 2026-11-04 02:39

[original source ↗](https://www.reddit.com/r/StableDiffusion/comments/1xmlzhi/best_way_to_combine_pose_identity_clothing_and/)

## retained text



```
I'm building a ComfyUI workflow using Krea 2 Turbo, a trained identity LoRA, clothing reference, background reference and DWPose/OpenPose LoRA.

I want to preserve the exact face, outfit and room while following a completely different pose reference. Currently, pose adherence is inconsistent and the clothing/background details also change.

Would you recommend a **two-pass workflow** (first generate the pose and composition, then refine identity, clothing and background), or is there a better single-pass conditioning method in Krea 2?

What's the most reliable approach, and what would you recommend I change?
```



document fetched 2026-11-04 00:31 · canonical https://reddit.com/r/StableDiffusion/comments/1xmlzhi/best_way_to_combine_pose_identity_clothing_and
