Controlnet Pose

Routes pose-conditioned generation across three RunComfy endpoints: Kling Motion Control for transferring a reference video's movement onto a target character, Wan 2-2 for audio-driven stylized animation, and Z-Image ControlNet LoRA for single-image pose conditioning from OpenPose skeletons or depth maps. The routing is automatic based on whether you're working with video or stills and photoreal versus stylized output. It's a wrapper around the runcomfy CLI, so you still need to figure out which control image type your LoRA expects and generate pose skeletons upstream. For multi-condition stacks like pose plus depth plus canny, you'll need to drop into ComfyUI workflows manually.