controlnet-pose

What problem does it solve?

ControlNet pose references are hard to translate into consistent image or video generation without manually choosing the right model endpoint and workflow.

Core Features & Use Cases

• Pose-conditioned routing: Routes automatically between video pose-transfer and image pose-conditioned generation based on whether the user provides a reference video vs a pose/control image.
• Multiple supported pose modalities: Handles motion/pose transfer using Kling motion control tiers and image conditioning using Z-Image ControlNet LoRA with control images like OpenPose skeletons, DWPose, canny edges, or depth maps.
• Practical outputs for creatives: Enables transferring a reference performance’s motion/blocking onto a target character or locking a character to a specific stance for photoreal or stylized results.

Quick Start

Use the controlnet-pose skill to generate pose-conditioned results by providing either a reference_video_url and character_image_url for motion transfer, or a prompt plus a control_image_url for control-image conditioning.