# Warlock-Studio
### _AI Media Enhancement Suite_
[](https://github.com/Ivan-Ayub97/Warlock-Studio/releases)
[](https://github.com/Ivan-Ayub97/Warlock-Studio/releases/tag/4.1)
[](LICENSE)
[](https://github.com/Ivan-Ayub97/Warlock-Studio/releases)
_Transform your media with cutting-edge AI technology_
---
**Warlock-Studio** is an open-source desktop application for **Windows** that integrates state-of-the-art AI models for video and image enhancement.
Inspired by [Djdefrag](https://github.com/Djdefrag) tools like **QualityScaler** and **FluidFrames**, this suite offers a unified, high-performance interface for **upscaling, restoration, and frame interpolation**.
Version **4.1** introduces improved GPU utilization, compatibility fixes, and optimized model loading for a faster, more stable experience.
---
## π₯ Download Installer (v4.1)
Get the latest stable release:
---
## β¨ Key Features
- **AI Upscaling & Restoration**
Real-ESRGAN, BSRGAN, IRCNN for denoising, upscaling, and detail recovery.
- **Face Restoration (GFPGAN)**
Revive blurry or low-quality portraits in photos and videos.
- **Frame Interpolation (RIFE)**
Generate **2Γ, 4Γ, 8Γ** smoother motion or slow-motion.
- **Modern UI**
Redesigned in v4.0 for an intuitive, streamlined experience.
- **Batch Processing**
Handle multiple media files simultaneously.
- **Custom Workflows**
Full control over models, resolution, output format, and quality.
- **Open-Source & Extensible**
Licensed under MIT, with a modular architecture for contributors.
---
## π Whatβs New in v4.1
- π§ Removed outdated SuperResolution-10 model.
- β
Robust ONNX loading & GPU acceleration.
- β
Fixed import errors & improved type annotations.
- β
Enhanced error handling with graceful fallbacks.
- π’ Better GPU utilization & resource management.
- π Compatibility fixes for NumPy & OpenCV.
- π¦ Stability improvements & refined memory usage.
- β
Improved startup reliability & user notifications.
---
## π Smart Model Distribution System (v4.0+)
### π― Lightweight Installation
- Installer reduced from **1.4GB β 450MB** (β68%).
- Models (~400MB) download automatically on first launch.
- Bandwidth-friendly setup.
### π‘οΈ Reliability
- **Integrity checks** on downloaded models.
- **Graceful degradation** if models are missing.
- **Offline support** for manual model placement.
---
## πΌοΈ Interface Previews
### Main Window

### RIFE Options

---
## π How to Use
1. Run as **Administrator** (recommended).
2. **Load Media**: Import images or videos.
3. **Configure Settings**:
- Choose AI model (Real-ESRGAN, GFPGAN, etc.)
- Set resolution, format, interpolation, etc.
4. **Start Processing** with **"Make Magic"**.
5. Retrieve results from the output folder.
---
## πΌοΈ Quality Comparison
Enhanced image using **BSRGANx2**

---
## π Model Comparison
| Model File | Use Case | Speed | Quality | Notes |
|--------------------------|------------------------------------------|---------|---------|-------|
| **GFPGANv1.4** | Face restoration | High | High | Great for blurry faces |
| **BSRGANx2** | 2Γ upscale + denoising | Medium | Very High | For lightly degraded images |
| **BSRGANx4** | 4Γ upscale + denoising | Low | Very High | For heavily degraded media |
| **RIFE** | Smooth frame interpolation | High | High | Best quality for motion |
| **RIFE-Lite** | Faster interpolation | Very High | Medium | Lightweight alternative |
| **RealESRGANx4** | General 4Γ upscale | Medium | High | Great all-rounder |
| **RealESRNetx4** | Subtle restoration without oversharpen | Medium | High | Preserves natural look |
| **RealSRx4_Anime** | Anime / line-art upscale | Medium | High | Clean edges for 2D art |
| **IRCNN_L** | Light denoising | High | Medium | Mild artifact removal |
| **IRCNN_M** | Medium denoising | High | Medium | Stronger artifact cleanup |
---
## βοΈ Installation
1. **Download installer** (links above).
2. **Run setup** and follow steps.
3. Launch from Start Menu / Desktop shortcut.
Warlock-Studio is packaged with **PyInstaller** + **Inno Setup**.
### Installer Previews



---
## π₯οΈ System Requirements
- **OS:** Windows 10+ (64-bit)
- **RAM:** 8GB+ recommended
- **GPU:** NVIDIA or DirectML-compatible GPU recommended
- **Storage:** Enough for input + processed media
---
## π Development Status (v4.1-08.01)
| Component | Status | Notes |
|-----------------------------|------------|-------|
| Upscaling Models | π’ Stable | VRAM recovery integrated |
| Optimized Model Suite | π’ Enhanced | Streamlined & reliable |
| Face Restoration (GFPGAN) | π’ Stable | High-quality face fix |
| Frame Interpolation (RIFE) | π’ Stable | Smooth motion, slow-mo |
| Batch Processing | π’ Stable | Improved error handling |
| User Interface (UI/UX) | π’ Refined | Clean, integrated models |
| GPU Management | π’ Enhanced | Robust ONNX + fallbacks |
| Code Quality | π’ Improved | Refactored & type-safe |
| Installer & Packaging | π’ Stable | Seamless setup |
---
## π Project Structure
```
Warlock-Studio/
βββ AI-onnx/ # Pre-trained ONNX models for AI processing
β βββ BSRGANx2_fp16.onnx
β βββ BSRGANx4_fp16.onnx
β βββ GFPGANv1.4.fp16.onnx
β βββ IRCNN_Lx1_fp16.onnx
β βββ IRCNN_Mx1_fp16.onnx
β βββ RealESR_Animex4_fp16.onnx
β βββ RealESR_Gx4_fp16.onnx
β βββ RealESRGANx4_fp16.onnx
β βββ RealESRNetx4_fp16.onnx
β βββ RealSRx4_Anime_fp16.onnx
β βββ RIFE_fp32.onnx
β βββ RIFE_Lite_fp32.onnx
β
βββ Assets/ # App assets and third-party binaries
β βββ banner.png
β βββ clear_icon.png
β βββ exiftool.exe
β βββ ffmpeg.exe
β βββ info_icon.png
β βββ logo.ico
β βββ logo.png
β βββ stop_icon.png
β βββ upscale_icon.png
β βββ wizard-image.bmp
β βββ wizard-small.bmp
β
βββ rsc/ # UI images, workflows, and branding
β βββ badge-color.png
β βββ Capture.png
β βββ CaptureRIFE.png
β βββ google_drive-logo.png
β βββ WorkflowBSRGAN.png
β βββ WorkflowIRCNN.png
β βββ WorkflowRealESRGAN.png
β βββ WorkflowRIFE.png
β βββ Installation_window2.png
β
βββ Manual/ # LaTeX sources and generated manuals
β βββ Manual_EN.tex
β βββ Manual_ES.tex
β βββ Manual_EN.pdf
β βββ Manual_ES.pdf
β
βββ Warlock-Studio.py # Main application script
βββ Warlock-Studio.spec # PyInstaller spec file
βββ Setup.iss # Inno Setup installer script
β
βββ README.md # Project overview (this file)
βββ requirements.txt
βββ CHANGELOG.md # Version history
βββ LICENSE # License info (standard)
βββ License.txt
βββ NOTICE.md # Notices and attributions
βββ CODE_OF_CONDUCT.md # Contributor behavior guidelines
βββ CONTRIBUTING.md # Contribution guide
βββ SECURITY.md # Security reporting policies
```
---
## π Integrated Technologies & Licenses
| Technology | License | Author | Source |
|---------------|------------|--------|--------|
| Real-ESRGAN | BSD/Apache | [Xintao Wang](https://github.com/xinntao) | [GitHub](https://github.com/xinntao/Real-ESRGAN) |
| GFPGAN | Apache 2.0 | [TencentARC](https://github.com/TencentARC) | [GitHub](https://github.com/TencentARC/GFPGAN) |
| RIFE | Apache 2.0 | [hzwer](https://github.com/hzwer) | [GitHub](https://github.com/megvii-research/ECCV2022-RIFE) |
| BSRGAN | Apache 2.0 | [Kai Zhang](https://github.com/cszn) | [GitHub](https://github.com/cszn/BSRGAN) |
| IRCNN | BSD/Mixed | [Kai Zhang](https://github.com/cszn) | [GitHub](https://github.com/cszn/IRCNN) |
| Anime4K | MIT | [bloc97](https://github.com/bloc97) | [GitHub](https://github.com/bloc97/Anime4K) |
| ONNX Runtime | MIT | [Microsoft](https://github.com/microsoft) | [GitHub](https://github.com/microsoft/onnxruntime) |
| PyTorch | BSD-3 | [Meta AI](https://pytorch.org/) | [GitHub](https://github.com/pytorch/pytorch) |
| FFmpeg | LGPL/GPL | [FFmpeg Team](https://ffmpeg.org) | [Site](https://ffmpeg.org) |
| ExifTool | Artistic | [Phil Harvey](https://exiftool.org/) | [Site](https://exiftool.org/) |
| PyInstaller | GPLv2+ | [Team](https://github.com/pyinstaller) | [GitHub](https://github.com/pyinstaller/pyinstaller) |
| Inno Setup | Custom | [Jordan Russell](http://www.jrsoftware.org/) | [Site](http://www.jrsoftware.org/isinfo.php) |
---
## π€ Contributions
We welcome community contributions!
1. **Fork** the repo
2. **Create a branch** for your feature/fix
3. **Submit a Pull Request** with details
π§ Contact: **[negroayub97@gmail.com](mailto:negroayub97@gmail.com)**
---
## π License
Β© 2025 IvΓ‘n Eduardo Chavez Ayub
Licensed under **MIT**. Additional terms in `NOTICE.md`.