wtf is hy-motion_extensions_text-to-motion?

This project is a beta extension for ComfyUI, a visual tool for building AI image and video pipelines. It takes human motion generated by HY-Motion, an AI model that turns text descriptions into movement, and applies that motion to a Mixamo rigged character. Mixamo is a common source of ready-made 3D character models with a standard skeleton, and this extension retargets the motion onto a T-pose FBX file of one of those characters using Blender running in the background. It is important to understand that this is only the retargeting step. The extension does not include the HY-Motion model itself, or any character files or textures. Users need to already have ComfyUI, the separate ComfyUI-HY-Motion1 node with the HY-Motion 1.0 model, Blender, and a Mixamo rigged T-pose FBX character before this tool becomes useful. The author confirms this works on Windows 11 with ComfyUI, Blender 5.1, and HY-Motion 1.0, tested with shorter-proportioned Mixamo characters. It only supports humanoid FBX files using Mixamo's standard bone naming, not VRM files, custom skeletons, non-human characters, or face and finger animation. Some foot sliding or limb overlap issues can still happen depending on the character's body proportions. Because the tool runs a local Blender script, the author warns to only use trusted FBX files. Hardware needs are significant: HY-Motion officially wants at least 26GB of video memory, though the author got a short 3 second test working on a 24GB card, and recommends 32GB or more of video memory, 64GB of system memory, and 100GB of free disk space for reliable use. Setup involves cloning the repository into ComfyUI's custom_nodes folder, installing Blender, placing a T-pose character file in the input folder, and loading one of two included workflow files, one for a local Windows setup and one for Google Colab. The custom node code itself is under the MIT license, while the HY-Motion model and any character assets remain under their own separate licenses.