anymove skill
anymove skills guide Claude Code: an animator places the creature’s joints, then the agent prepares motion capture and GLB export.
- 543 ★ on the repository
- Last commit
- Apache-2.0
Extended access
Why this level
- The agent runs processing and training, writes project files and can call paid fal APIs; batch spending must be agreed first.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/anymoveDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The agent runs processing and training, writes project files and can call paid fal APIs; batch spending must be agreed first.
This is third-party code. Review the repository files before installing.
What it does
anymove is a Claude Code skill collection for capturing creature motion from AI video. Each frame must contain matching front and side views. A model locates joints in both views, and fitting tools align the skeleton with the resulting points. For a new creature, the user manually places joints on grid views and previews the skeleton. The agent then trains the joint-detection model, prepares a video and fits its motion to the rig. The animator compares the render with the source clip before GLB export. Each creature keeps its files in a separate directory.
Who it is for. 3D animators experimenting with AI-video motion for dragons, spiders and other non-humanoid rigs.
Good fit when
- A rigged character needs a short gesture from a text description.
- A synchronized two-view clip is ready to fit onto a skeleton.
- You want to test the workflow with the provided dragon before using your own creature.
- A failed capture needs comparison between the resulting render and source motion.
Not a fit when
- You expect motion capture from arbitrary single-camera footage.
- The task requires reliable eight-legged walking or lying-down motion.
- There is no opportunity to place joints manually for a new creature.
Example request
Test anymove’s prepared dragon-roar fit on CPU and show fitting error without paid generation.Limitations
Training the joint-detection model requires an NVIDIA CUDA GPU with 24 GB VRAM; the prepared fitting example runs on CPU. The full workflow uses Python 3.11 and Blender 5.x, so prepare the environment before agent work. New fal clips require a key and paid requests; approve the batch cost first. Motion in the two AI-video views can drift apart. Compare detected joints and the render before export, especially for creatures with many limbs.
How to disable. Disable anymove in Claude Code and stop active training or rendering jobs. Generated files and downloaded models remain in the project directory.
FAQ
Which commands add the skills?
Run `/plugin marketplace add Patch-gg/anymove`, then `/plugin install anymove@anymove` in Claude Code.
Can I try it without a GPU?
Yes. The repository includes detected joints for a dragon roar. That fitting example runs on CPU without Blender or new downloads.
Where does my own mesh go?
Place it at `assets/<name>/model.glb`, then follow scale, orientation and rig preparation.
Who places the joints?
The user manually marks joints on grid views. Before training, open the preview and inspect the bone positions.
Why paint limbs in different colors?
Distinct colors help distinguish left and right limbs in video. Check generated clips for color swaps anyway.
Discussion
No questions or reviews yet.
Related
SkillsA Claude Code skill that builds motion graphics in code: 21 ready looks, five questions, three approval stills and an MP4 render.
A Claude Code skill that builds a game-ready 3D asset in Blender from concept art and checks each phase by measurement.
A Claude Code and Codex skill that turns a topic into a narrated motion-graphics explainer video on Remotion with subtitles and chapters
A video editing skill for Claude Code and Codex: the agent cuts pauses and filler words, grades color, burns subtitles and adds animated overlays
A Claude skill for Spine 2D animation: it positions character parts, builds the skeleton JSON and prepares an HTML preview
A Claude Code skill for short animated videos drawn in code: seven styles, story and storyboard sign-off, voice-over and an MP4 render
README in short
The collection includes a separate skill for training and testing a MiniMax H3 LoRA. The code uses Apache-2.0, while MiniMax weights and derived LoRAs have separate terms. The repository license does not automatically grant permission to use them.
README badge
Original SKILL.md
--- name: anymove description: Motion capture for any creature from one AI video, with the anymove repo. A video model films a rigged 3D creature from two cameras in one frame, a DINOv2 joint reader trained on that creature finds every joint in both views, and the rig is fitted to them and exported as a GLB. Use when someone wants to animate a non-human creature (dragon, spider, wolf, monster) from a text prompt or an AI video, asks how anymove works, or needs to know which step comes next; it routes to anymove-rig, anymove-film, anymove-reader, anymove-capture and anymove-lora. license: Apache-2.0 compatibility: Needs an NVIDIA GPU with CUDA (24 GB to train a reader at batch 20), Blender 5.x, Python 3.11, and a fal.ai key for video generation. --- # anymove Animate any rigged creature from one AI video. The trick is the frame: one image holds two orthographic views of the creature at the same scale, front on the left and side on the right, sharing the vertical axis. A joint sits on the same pixel row in both halves, so the front view gives left-right and height, the side view gives front-back and height, and together they pin the joint in 3D. The method is @CanItCode's; the LoRA step is @Graalitoo's idea. ## The pipeline | step | once per | what happens | skill | |---|---|---|---| | 1. mesh | creature | a textured 3D model (Tripo H3.1 from one image, or your own GLB), facing -Y, real size | anymove-rig | | 2. skeleton | creature | joints placed in `assets/<c>/rig.json`, rig built and skinned in Blender | anymove-rig | | 3. limb colours | creature | each limb painted its own colour so left and right never swap | anymove-rig | | 4. two cameras | move | the painted reference frame plus a strict prompt go to MiniMax H3 Max on fal | anymove-film | | 5. teach a reader | creature | 2,200 random poses rendered with known joints; DINOv2 learns this creature | anymove-reader | | 6. read every joint | clip | the reader finds every joint in both views, every frame | anymove-capture | | 7. two views, 3D | clip | shared heights tie the views; the pair gives each joint in 3D | anymove-capture | | 8. the rig follows | clip | forward kinematics in torch fitted to both views; GLB out | anymove-capture | | optional | creature | a MiniMax H3 LoRA that keeps the two-camera layout and colours more reliably | anymove-lora | ## Before anything 1. Work in an anymove checkout. If the working directory has no `tools/capture.sh`, clone it: `git clone https://github.com/Patch-gg/anymove && cd anymove`. 2. Set up once: `python3.11 -m venv .venv`, then `.venv/bin/pip install torch==2.11.0 --index-url https://download.pytorch.org/whl/cu128` and `.venv/bin/pip install -r requirements.txt`. Blender 5.x must be on PATH, or set `BLENDER=/path/to/blender`. 3. For video generation and LoRA training, put `FAL_KEY=...` in `.env` (git ignores it) or export it. Never print it. 4. Every command takes a creature name; its config is `assets/<c>/creature.json` and every output lands in `runs/<c>/` (reference frame, dataset, `reader.pt`, videos, clips). The scripts set `MOCAP_CREATURE` for you; when you run a `python -m mocap...` module by hand, set `MOCAP_CREATURE=assets/<c>` first, or the wolf's layout applies.