Krea2 Multi-Reference Image Editing Workflow

This workflow is for Krea2 multi-reference image editing. It is designed for users who want to edit or rebuild an image with multiple visual references instead of relying on text alone. You can use different reference images to guide character identity, clothing, pose, background, mood, color, or style, then use the prompt to define the final edited result.

The workflow uses Krea2 Turbo FP8, Qwen3-VL as the Krea2 text encoder, Qwen Image VAE, multiple LoadImage nodes, and Krea2EditRebalance. The Krea2EditRebalance node is the key editing-control part of the graph. It helps the workflow understand how the references should influence the final image while still leaving space for prompt-driven transformation.

This workflow is useful for character redesign, background replacement, style fusion, multi-image reference editing, product-style visual transformation, poster design, and fantasy character creation. It is more flexible than single-reference style transfer because it can use several images at once, but it also requires clearer prompting. The prompt should state what comes from each reference and what the final image should become.

Main features:

• Krea2 multi-reference image editing

• Multiple image input support

• Krea2EditRebalance control

• Krea2 Turbo FP8 model route

• Qwen3-VL Krea2 text encoder

• Qwen Image VAE decoding

• Reference-based character and style editing

• 1:1 square output route

• 8-step sampler setup

• Suitable for image editing, redesign, and reference fusion

Suggested workflow:

Upload clear reference images and define their roles. For example: keep the character from image 1, use the background from image 2, borrow the color mood from image 3, and follow the prompt for the final design. Do not leave the references ambiguous. If the result becomes chaotic, reduce the number of references or make the prompt more explicit. This workflow works best when each image has a specific job.

⚙️ RunningHub Workflow

Try the workflow online right now — no installation required.
👉 Workflow: https://www.runninghub.ai/post/2073314422151540738?inviteCode=rh-v1111

If the results meet your expectations, you can later deploy it locally for customization.

🎁 Fan Benefits: Register to get 1000 points + daily login 100 points — enjoy 4090 performance and 48 GB super power!

📺 Bilibili Updates (Mainland China & Asia-Pacific)

If you’re in the Asia-Pacific region, you can watch the video below to see the workflow demonstration and creative breakdown.
📺 Bilibili Video: https://www.bilibili.com/video/BV1U1TS6gEVe/

☕ Support Me on Ko-fi

If you find my content helpful and want to support future creations, you can buy me a coffee ☕.
Every bit of support helps me keep creating — just like a spark that can ignite a blazing flame.
👉 Ko-fi: https://ko-fi.com/aiksk

💼 Business Contact

For collaboration or inquiries, please contact aiksk95 on WeChat.

⚙️打开下方链接即可在线体验，无需安装。
👉 工作流： https://www.runninghub.ai/post/2073314422151540738?inviteCode=rh-v1111
如果觉得效果理想，你也可以在本地进行自定义部署。

🎁 粉丝福利： 注册即送 1000 积分，每日登录 100 积分，畅玩 4090 体验 48 G 超级性能！

📺 Bilibili 更新（中国大陆及南亚太地区）

如果你在中国大陆或南亚太地区，可以通过下方视频查看该工作流的实测效果与构思讲解。
📺 B站视频： https://www.bilibili.com/video/BV1U1TS6gEVe/

我会在 夸克网盘 持续更新模型资源：
👉 https://pan.quark.cn/s/20c6f6f8d87b
这些资源主要面向本地用户，方便进行创作与学习。