ShapeEmote — Facial Shape Key Tools

Give your character a wider range of expression.

ShapeEmote helps VTuber and avatar artists turn a prepared face mesh into editable facial shape keys. Register your face parts and eyelid/mouth guides, generate starting shapes, and refine each expression to suit your character.

REGISTER. MIX. MAKE IT YOURS.

• One-click object detection: select the face and detect the remaining parts. Existing assignments are kept. Check and correct automatic matches when needed.

• 13 expression presets: six emotions plus seven additional expressions. Happy, Angry, Sad, Relaxed, Surprised, Laugh, Smirk L/R, Scared, Open Grin, Lip Bite and Tongue Out L/R. More expressions are planned.

• 15 VRChat visemes, including the five vowels A / I / U / E / O.

• 52 Perfect Sync shape keys: 29 mouth-region and 23 eye/brow-region shapes. These are editable shapes, not 52 finished emotion presets. Toon-style noseSneer moves the inner brows rather than deforming the nose.

• Blinks, smiling eyes, winks and four-direction gaze shapes, with Eyes + Eyelids or Eyelids Only modes.

• Preview matching keys across registered facial parts without joining the meshes. Combine strengths, then register the mix as a custom expression.

• Adjust individual expressions, refine shape-key vertices, mirror expressions, smooth shape-key deltas and reuse settings or same-topology setup profiles.

SETUP → CREATE → FINISH

1. Select the face, register the facial parts and eyelid/mouth guides, and check the affected regions.

2. Choose a naming format, generate the categories you need, and preview the results.

3. Adjust and refine the shapes for your character, or combine them into custom expressions.

FOUR NAMING FORMATS

Standard / VRChat / VRM 0.x / VRM 1.0.

These options set generated key names. Exporter bindings, tracking applications and avatar configuration are separate; the addon does not automatically produce an export-ready avatar.

BEFORE YOU BUY

• Requires Blender 4.5 or later. Japanese and English UI. No extra addon, external Python library or internet connection is required during use. macOS and Linux have not been tested.

• Intended for neutral stylized humanoid faces with identifiable eyelid and mouth contours. Automatic guide detection uses suitable closed seam or sharp-edge contours; manual correction is available.

• Supports separate facial objects or registered part ranges within an integrated mesh. Shared-mesh ranges must be complete and non-overlapping. Keep topology stable after setup.

• Generated shapes are editable starting points. Review and finish the results for your model, including clearance around eyelashes, lips and teeth. Automatic detection is not guaranteed to register every part correctly.

• Regeneration may affect existing same-name keys. Keep a saved copy of your model before editing.

PACKAGE AND SUPPORT

Version 1.0.0. Python addon licensed under GPL-3.0-or-later. English and Japanese written guides and a changelog are provided separately from the installable addon ZIP.

Character models, textures and demonstration scenes are not included. Preview images show an example model, including model-specific expression adjustments.

For product questions or reproducible errors, use Superhive product support and include your Blender version, addon version, steps to reproduce and error text.