Tools: How to Copy Any Pose in ComfyUI and Fix AI Skin

Set Up the Pose Recognition ## Extract Poses From Complex Images ## Add Cinematic Lighting Automatically This is a summarized guide. For the full JSON workflow and download files, check the original article on my site: https://aistudynow.com/how-to-copy-any-pose-without-vnccs-pose-studio-comfyui-workflow/ Manual 3D posing wastes time. You can extract poses directly from photos instead. This workflow auto-generates missing body parts and fixes plastic-looking skin. You control exactly what your character wears. The system offers a switch between GGUF and Safetensors models. You can also pick your image resolution instantly. Here is the exact process. The AI needs to understand the body position. You need a specific node for this. Add the following node created by aistudynow: Open the node settings. Select this exact preset: This preset writes a perfect text prompt describing the body position. The model now knows what to do. Next, you need a specific LoRA. Download it and save it to your LoRA folder. Make sure it stays enabled. If you turn it off, you will get random backgrounds. VNCCS PoseStudio V5 LoRA Let us test a difficult pose. We will use an image of a hand showing a middle finger. Most AI models fail at this. Enable "Camera 2" in the workflow. This camera processes external images. Upload your reference picture. Enter this exact trigger text into your prompt: Draw character from image 2 Run the workflow. The AI captures the posture perfectly. Fix Plastic Skin and Enhance Details The initial result might look like cheap plastic. The Qwen 2511 model often causes this issue. We will add details to fix it. Go to the detailer group: Select the face target using this detector: Run the prompt again. The workflow crops and enhances the face. It looks better now. It still has an artificial look. Go to the end of the workflow. Enable the final detail group: Run it one last time. The plastic skin vanishes. The character looks highly realistic. The exact hand pose remains intact. Generate Missing Body Parts Heavy detailers consume high VRAM. They can slow down your PC. Disable the add details and Z-Image groups for this next step. Load a half-body reference image. The AI detects the missing lower half. It guesses the rest of the body automatically. We want full control over the clothing. Go to your text prompt. Add this simple instruction: Run the workflow. The AI generates a full body. It dresses the woman in jeans perfectly. Build Custom Poses From Scratch You can also build custom poses using the VNCCS Pose Studio. Turn the 3D studio back on. Select the switch to enable "Camera 1". Look at the joints on the 3D dummy. Click the tiny yellow dot on any joint. Colored 3D circles will appear. Click and hold one circle. Drag your mouse. You can move the arm anywhere. Bend the elbow to create a perfect position. Go to your text prompt. Delete the clothing text. Leave only the magic trigger: Draw character from image 2 Run it. The AI builds your exact custom pose. You do not need to type long prompts for shadows. We control lighting directly in the node. Find the lighting hook on the Pose Studio. Connect this wire directly to your text box: Lighting Prompt -> Text 1 Open the lighting menu inside the 3D studio. Select a light. Change its direction. Drag it next to your 3D dummy. Run the generation again. The system reads your 3D lights. It adds beautiful and dramatic lighting to the final image. You get perfect shadows automatically. Templates let you quickly answer FAQs or store snippets for re-use. Are you sure you want to hide this comment? It will become hidden in your post, but will still be visible via the comment's permalink. Hide child comments as well For further actions, you may consider blocking this person and/or reporting abuse