This workflow is prepared for creators who want a cleaner Anima Aesthetic 1.1 image without rebuilding a high-resolution pass by hand. It focuses on a controlled 1.5x latent upscale workflow: generate the composition first, then pass the latent through a second refinement stage so the final image keeps the original character shape while gaining sharper line work, cleaner eyes, and more stable surface detail. It is a practical setup for anime illustration, character posters, vertical covers, and thumbnail tests where you need more polish than a single txt2img pass can usually provide.

The workflow uses a Qwen Image text encoder and Qwen Image HDR VAE, with an Anima Aesthetic V1.1 model selection path in the graph. The main generation chain includes a full first sampling pass and a lower-denoise refinement pass after latent scaling. The inspected graph shows an er_sde/simple sampler route, a 34-step base pass, an 18-step refinement pass, CFG around 4.5, and a 1.5x LatentUpscaleBy stage. A vertical 9:16 style resolution selector is present, so this is especially suitable for mobile-friendly artwork and cover-format compositions. The description intentionally avoids claiming 4K output because the confirmed upscale stage is latent 1.5x, not a full pixel super-resolution pipeline.

Main features:

- Anima Aesthetic 1.1 oriented image refinement workflow
- 1.5x latent upscale for higher-detail outputs
- Two-stage sampling structure for composition then polish
- Qwen Image CLIP text encoder
- Qwen Image HDR VAE decode path
- er_sde/simple sampling route
- 34-step first pass and 18-step refinement pass observed
- CFG around 4.5 for balanced prompt adherence
- Vertical composition support through the resolution selector
- Useful for character covers, posters, and preview images
- Conservative setup without advertising inactive bypassed nodes
- Designed for direct online testing through RunningHub

Suggested workflow:

Start with a clear character or scene prompt and keep the first run simple. Use the base pass to lock the pose, silhouette, and camera framing. Once the composition is stable, use the 1.5x latent refinement stage to recover details in the face, hair, costume edges, and background lighting. If the image starts drifting during refinement, lower the denoise of the second stage rather than changing the prompt too aggressively.

⚙️ RunningHub Workflow

Try the workflow online right now — no installation required.
👉 Workflow: https://www.runninghub.ai/post/2078721287299805186?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/2078721287299805186?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。