This workflow is an experimental Anima-Edit testing setup for comparing LoRA-based editing and IC/reference-style methods inside ComfyUI. It is not written as a one-click beauty preset. The value is that it keeps several edit routes in one graph so you can study how an image responds to inpaint conditioning, reference latent injection, Anima LLLite inpainting, and dedicated edit LoRAs. It is useful for creators who want to test identity retention, outfit changes, controlled variations, or character edits without rebuilding the graph from zero.

The inspected workflow uses Qwen Image CLIP, Qwen Image HDR VAE, and an enhanced diffusion loading path. The active LoRA entries include `AnimaEditV1.safetensors` and `lora_edit_ZeroTwo.safetensors`; an older turbo LoRA appears but is bypassed and should be treated as inactive. The control side includes InpaintModelConditioning, ApplyCosmosReferenceLatent nodes, and an Anima LLLite inpainting model. The sampling routes use 20-step `dpmpp_2m_sde_gpu` style passes with CFG around 5. Multiple SaveImage outputs are present for comparison, which makes the workflow especially useful when you want to see how each edit path behaves under similar prompt and input conditions.

Main features:

- Experimental Anima editing workflow
- Active AnimaEditV1 LoRA route
- Active lora_edit_ZeroTwo test route
- InpaintModelConditioning support
- Cosmos reference latent application nodes
- Anima LLLite inpainting control path
- Qwen Image CLIP text encoding
- Qwen Image HDR VAE route
- 20-step edit sampler passes observed
- Multiple output branches for comparison
- Useful for identity, outfit, and reference-guided edit tests
- Bypassed turbo LoRA is not described as active

Suggested workflow:

Start with one input image and one simple edit instruction. Compare the LoRA edit route and the reference/inpaint route before adding more variables. If identity drifts, reduce the strength of the edit prompt and rely more on the reference route. If the edit is too weak, increase the prompt clarity first, then adjust LoRA strength or denoise in small steps.

⚙️ RunningHub Workflow

Try the workflow online right now — no installation required.
👉 Workflow: https://www.runninghub.ai/post/2078721128689258497?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/BV1xJKB6vEfp/

☕ 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/2078721128689258497?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/BV1xJKB6vEfp/

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。