anymove

anymove converts AI-generated two-view creature videos into skeletal animations for a rigged 3D model

来源语言 · EN

产品介绍

anymove is an open-source creature motion-capture workflow from Patch. It uses a front-and-side AI video, a joint reader trained for the individual creature, and skeletal fitting to export animation as a GLB file. Creating new captures requires an NVIDIA GPU, Blender, Python and a fal API key. The software is free to use under Apache-2.0; AI video generation and any rented compute cost extra. Large gestures work better than eight-legged walking, lying down or fine foot contact, and a different creature needs its own reader.

来源语言 · EN
主要功能

anymove 能做什么

  1. 01Read creature joints from synchronized front and side video views
  2. 02Train a joint reader for an individual rigged creature
  3. 03Fit skeletal motion and export GLB animations
  4. 04Score takes and assemble animation packs
来源语言 · EN
差异特点

它有什么不同

The workflow handles nonhuman body plans using two views in a single generated clip. Author-reported costs and success rates depend on repeated takes; generated assets retain their model-specific terms.

来源语言 · EN
适合人群

谁值得考虑

  • Technical artists animating creatures
  • Game developers comfortable with local GPU workflows
来源语言 · EN
选择之前

需要核对的事项

  • 用一个真实任务试用,并对照自己的要求检查结果。
  • 核实服务会保存、共享或用于训练哪些数据。
  • 在官网确认当前价格、使用限额和取消条款。
官网核验日期

官网链接核查于 2026年11月16日;功能声明尚未经过独立测试。