This workflow is the lightweight silent version of the MiniMax H3 character replacement / reference editing route. It keeps the same core swap chain as the standard version but removes the dual-instance duplication, the face refine add-on, and the audio mix stage, so it is the fastest variant in the set for motion-first testing.

The inspected graph is a single-instance build: MiniMax H3 hybrid FL2VA/Ref2VA INT8 base model (minimax_h3_hybrid_fl2va_ref2va_zs05_b25-49_int8.safetensors), Qwen3-VL 32B NVFP4/AWQ text encoding, MiniMax H3 video VAE (INT8 ConvRot) and audio VAE (FP32), OpenVDN DMD8 execution planning with a custom sampler, Video Depth Anything ViT-S depth (518x1280 fp16, gray), SeC full-clip segmentation with feathered compositing, and the MiniMaxH3ReferenceToVideo control node (Picture 1 identity reference plus single reference video depth composite, 768x768, 362-frame setting). The final clip is still packaged at a unified 24 FPS with a safe audio channel; no audio mix stage is present, matching the silent positioning of this variant.

This package is meant for users who want the swap result without audio post-processing and with the smallest possible graph. It is useful for fast iteration on character identity and motion, and for testing on lighter setups before committing to the full standard or face refine variants.

Main features:

- Lightweight single-instance character replacement graph (38 nodes vs the dual-instance variants)
- Same core chain: MiniMax H3 hybrid INT8 base, Qwen3-VL 32B NVFP4/AWQ encoding, OpenVDN DMD8 sampling
- Video Depth Anything ViT-S whole-frame depth and SeC full-clip segmentation with feathered compositing
- No face refine add-on and no audio mix stage; unified 24 FPS packaging with a safe audio channel
- RGB + depth person display preview output for compositing checks
- Fastest route in the three-variant set for motion and identity iteration

Suggested workflow:

Use this version when you want to iterate quickly on the swap itself: same reference image and video as the other variants, shorter duration for tests, then review character identity, motion plausibility, and depth compositing edges. Once the swap behaves correctly, move to the standard or face refine variant for final renders with audio or face cleanup.

For fair comparisons across the three variants of this episode, run one short test first with identical inputs, then review identity stability, motion consistency, and background drift before scaling up to full-length renders.

RunningHub Workflow

Try the workflow online right now - no installation required.
Workflow: https://www.runninghub.ai/zh-cn/post/2106048566154698754?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/BV1VjaD6jEU3/

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.
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/zh-cn/post/2106048566154698754?inviteCode=rh-v1111

如果你觉得效果理想，也可以在本地进行自定义部署。

粉丝福利：注册即可领取 1000 积分，每日登录再领 100 积分，体验 4090 和 48GB 大显存性能。

B站视频（中国大陆及亚太地区）

如果你在中国大陆或亚太地区，可以通过下面的视频查看工作流的实测效果与创作思路。
B站视频：https://www.bilibili.com/video/BV1VjaD6jEU3/

我会在夸克网盘持续更新模型资源：
https://pan.quark.cn/s/9ec095b95838