RepositoriesOctoRig Studio

Blender + UniRig toolkit · early development

OctoRig Studio

Rig a 3D character, animate its body and face, and check the result. On your machine.

Local tools that prepare a character rig with an adapted UniRig, build editable body and facial animation with Blender scripts, retarget existing clips and let you inspect the result in a GLB viewer. No account, and no models or checkpoints bundled.

  • Python 3.11+
  • Blender
  • UniRig (CUDA, PyTorch)
  • MIT · GPL-3.0 modules
Body: skeleton and skinning generated with OctoRig and UniRig, animated in Blender with custom scripts. Face: a separate bust with a custom jaw rig. Modelling, materials, lighting and rendering were separate steps; this is not a one-click result.

What's included

From a static model to an editable performance

Rigging in stages

Prepare a static GLB, predict a skeleton with UniRig, review it, then infer the skin weights. Each job records its paths and hashes, and the skeleton is approved by hash before skinning.

Body animation

Map bones explicitly, describe editable poses, bake named clips, and export GLB and Blender files.

Facial controls

Configure jaw landmarks and shape-key displacements, then create editable facial controls and animation in Blender.

Retargeting

Transfer existing clips with an explicit bone mapping between compatible rigs.

Inspect and view

Inspect and compare GLBs from the command line, and play existing animations in a local viewer that runs in your browser and rejects external resources.

Early development

What it doesn't do (yet)

  • The Blender tools have been tested on synthetic rigs. They still need artistic review on real characters.
  • It doesn't animate every type of object automatically, create lip-sync from audio, or guarantee good deformation and ground contact.
  • Profiles and facial landmarks have to be configured for each model.
  • Automatic rigging needs a separate UniRig environment (CUDA, PyTorch) and checkpoints you download yourself.

Install

The Python tools

From a checkout

git clone https://github.com/Octonove/octorig-studio.git
cd octorig-studio
python -m venv .venv
python -m pip install -e .
python -m octorig inspect examples/wave.glb

Activate .venv before the pip line. The checkout includes the viewer, the UniRig source and the docs; an installed wheel doesn't.

You need

  • Python 3.11 or later, with NumPy
  • Blender, for authoring
  • For automatic rigging: a UniRig environment and its skeleton and skin checkpoints

Original code is MIT; selected Blender and animation modules are GPL-3.0-or-later. The orc clips are © 2026 Octonove, shown for viewing only.

The code, the issues and the history are on GitHub.

Star it if it helps; open an issue if it breaks.

View on GitHub