This workflow is designed for Z-Image Turbo 2602 ControlNet-guided image generation. Its main purpose is to let creators use a reference image as a structural control source, extract pose or layout information through a preprocessor, and then generate a new image with Z-Image Turbo while keeping the composition more stable than pure text-to-image generation.
The workflow uses z_image_turbo_bf16.safetensors as the main generation model, qwen_3_4b.safetensors as the text encoder, and ae.safetensors as the VAE. It also applies the Z-Image Fun Distill 4-step 2602 LoRA route, making the workflow suitable for faster Z-Image Turbo generation while still keeping ControlNet guidance active. The ControlNet part uses Z-Image-Turbo-Fun-Controlnet-Union-2.1-2601-8steps.safetensors through ModelPatchLoader, then sends the control patch into ZImageFunControlnet.
The main control logic comes from the uploaded reference image. The image is processed by AIO_Preprocessor with DWPreprocessor, which extracts pose / skeleton-style guidance from the input. This guidance image is then used by the Z-Image ControlNet route to control the final image structure. In practical terms, this means the output can follow the body pose, gesture, camera framing, or general composition of the reference image, while the prompt defines the new subject, style, clothing, lighting, atmosphere, and visual direction.
This workflow is especially useful when text prompts alone are not enough. If you want a character to hold a specific pose, follow a reference composition, or maintain a clearer body structure, ControlNet guidance gives the model a stronger spatial anchor. Z-Image Turbo then handles the actual image generation, making the workflow faster and more suitable for online testing, RunningHub publishing, Civitai examples, and rapid prompt iteration.
The sampling structure uses ModelSamplingAuraFlow, CFGGuider, BasicScheduler, SamplerEulerAncestral, RandomNoise, and SamplerCustomAdvanced. This gives the workflow a more modular generation pipeline than a basic KSampler-only setup. The scheduler is configured around a short step count, which fits the Turbo and distill-style design. The result is decoded through VAEDecode and previewed directly in the workflow.
The workflow also keeps the prompt area flexible. Users can write detailed prompts for anime characters, fantasy portraits, fashion photography, cinematic poster scenes, creature designs, or stylized concept art. The example prompt in the graph includes fantasy character elements, close-up composition, detailed lighting, spirit cats, surreal floating objects, and a colorful starry background, showing that the workflow is not only for simple pose transfer but also for highly stylized visual generation.
In short, this is a fast Z-Image Turbo 2602 ControlNet workflow for creators who want stronger pose and composition control without losing the speed advantage of Turbo generation. If you want to see how the reference image is preprocessed, how the ControlNet patch is connected, and how Z-Image Turbo generates the final controlled output, watch the full tutorial from the YouTube link above.
⚙️ Try the Workflow Online
👉 Workflow: https://www.runninghub.ai/post/2029507207917735938?inviteCode=rh-v1111
Open the link above to run the workflow directly online and view the generation results in real time.
If the results meet your expectations, you can also deploy it locally for further customization.
🎁 Fan Benefits: Register now to get 1000 points, plus 100 daily login points — enjoy 4090-level performance and 48 GB of powerful compute!
📺 Bilibili Updates (Mainland China & Asia-Pacific)
If you are in Mainland China or the Asia-Pacific region, you can watch the video below for workflow demos and a detailed creative breakdown.
📺 Bilibili Video: https://www.bilibili.com/video/BV141Pkz4E95/
I will continue updating model resources on Quark Drive:
👉 https://pan.quark.cn/s/20c6f6f8d87b
These resources are mainly prepared for local users, making creation and learning more convenient.
⚙️ 在线体验工作流
👉 工作流： https://www.runninghub.ai/post/2029507207917735938?inviteCode=rh-v1111
打开上方链接即可直接运行该工作流，实时查看生成效果。
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/BV141Pkz4E95/
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。