Repository navigation
🚀 Greko Engine (Windows Preview)
Pre-releaseGreko Engine is a lightweight VRM rendering tool designed for developers, artists, and hobbyists who want to load and interact with VRoid / VRM characters outside of traditional VTubing software.
This Windows release focuses on providing a clean, programmable rendering pipeline for VRM models through a native C++ OpenGL backend with Python integration.
Unlike streaming-focused avatar tools, Greko is built as a developer sandbox and experimentation platform.
✨ What This Is
- A standalone VRM viewer + renderer
- A developer-friendly sandbox for experimenting with VRM characters
- A bridge between Python scripting and native C++ rendering
- A Windows-native
.pydpowered rendering runtime
❌ What This Is NOT
- Not a VTuber app (no streaming, face tracking, webcam integration)
- Not a full game engine (yet)
- Not a polished production tool (early alpha build)
🚀 Features
🧍 VRM / VRoid Model Support
- Loads
.vrmfiles directly - Full skeleton support
- Tested with 133–200 joints
- Compatible with older VRM 1.x models
- Morph targets / blendshapes fully working
🎭 Basic Character Animation Behaviors
- Blinking
- Breathing
- Mouth sequencing (phoneme-based)
- LookAt (bone-driven head tracking)
🎮 Interactive 3D Camera
- WASD movement
- EQ vertical movement
- Mouse look
- ESC mouse toggle
🧠 Python-Controlled Behavior System
- Extend character logic through Python scripts
- Runtime behavior injection
- Modular behavior pipeline
Example systems:
BlinkerBreatherMouthSequencerLookAt
⚡ Native C++ OpenGL Backend
- Hardware accelerated rendering
- Skeletal animation
- GPU morph blending
- Transparent / opaque render sorting
- OpenGL 4.3 backend
🔁 Hot-Swappable Native Core
- Replace the
.pydbackend to update engine features - Python-side systems remain unchanged
- Fast iteration workflow for development
📦 Included
launcher.py→ Windows GUI launchergreko_native.cp313-win_amd64.pyd→ Native rendering backend- glfw3dll.lib → for glfw3 rendering
- _internals → internal deps
- Python launcher + asset selector
- VRM model runtime loader
🛠️ How to Use
For Windows
-
Run the launcher:
- Download the Zip Folder and extract it in a separate standalone folder.
- Run the launcher/launcher.exe file.
- You can delete the engine_cache and config.json (They will regenerate when you run the program)
- Make sure you do not forget any of the files
-
In the UI:
- Select the
.pydfile - Select the folder containing
.vrmfiles - Choose a model (VRM1.x models only!)
- Click Run Engine
- Select the
🎯 Use Cases
- 🧪 Quickly testing VRM models
- 🎮 Character system prototyping
- 🤖 AI-driven avatar experiments
- 🧩 Building custom avatar tools
- 🧠 Future cognitive agent integration
🧪 Status
This build is functional and stable for testing, but still under active development.
Known limitations may include:
- early UI roughness
- limited asset tooling
- experimental behavior APIs
- performance tuning still in progress
🪟 Platform
- ✅ Windows (OpenGL 4.3+)
- ✅ Linux supported separately
- 🔄 Cross-platform parity achieved
🔮 Future Direction
Greko Engine is evolving toward:
- advanced character behavior systems
- AI-driven avatar control
- perception and memory systems
- modular runtime architecture
- interactive world simulation
This release focuses on the rendering + animation foundation.
🙌 Feedback
If you try this build:
- report bugs
- suggest features
- share ideas
- test different VRM models
Your feedback directly shapes the next release.
🔥 Closing Note
There are many tools for streaming VRM avatars.
Greko exists for something different:
Rendering, controlling, and experimenting with VRM characters freely on Windows.
This is the first major Windows-native alpha build.
More updates coming soon.