VoxCraftRecap

Notice: (The source code has been analyzed and reverse-engineered—if you encounter errors, fix them yourself.)

VoxCraftRecap is an all-in-one AI-powered Movie Review and Video Recap Generator. The application automates the end-to-end video creation pipeline: from subtitle extraction and AI scene analysis to Gemini-powered narrative script generation, multi-voice text-to-speech (TTS), smart audio/video synchronization, transformative fair-use visual filtering, and automated multi-threaded FFmpeg rendering.

💖 Support the Original Developer: Please consider supporting the original creator of VoxCraft by visiting the official website: https://voxcraftone.com/

---

---

✨ Key Features

• 🎙️ Automatic Subtitle Extraction (Faster-Whisper): Rapidly extracts SRT subtitles directly from source video files with optional GPU acceleration (CUDA).
• 🧠 AI Script Writing (Google AI Studio / Gemini API): Analyzes movie timelines and generates structured, dramatic recap scripts based on detailed scene breakdowns.
• 🗣️ Multi-Voice Text-to-Speech (TTS Engine):
  • Integrates TikTok TTS and Edge-TTS with expressively tuned voice storytellers.
  • Features proper name phonetization, custom replacement dictionaries, speech speed controls, and smart pause inserts.
• ⚡ Smart A/V Synchronization Engine:
  • Dynamically calculates video stretch factors (setpts) and audio background tempo (atempo) to ensure seamless alignment between narration and source clips without audio distortion or A/V drift.
• 🛡️ Fair-Use & Transformative Visual Effects:
  • Built-in Phantom Shift filter suite (GOP modification, subtle color grading, dynamic zoompan, noise injection, edge blur, custom logo/watermark overlays).
• 🎬 Hardware-Accelerated Video Rendering (FFmpeg):
  • Auto-detects available GPU encoders: NVIDIA NVENC, AMD AMF, Intel QuickSync (QSV), or falls back to optimized CPU libx264.
• 💻 Modern Dark Mode UI: Built with CustomTkinter for a clean, intuitive user experience.

---

📁 Repository Structure

VoxCraftRecap/
├── ai_movie_review.py          # Main GUI Application (CustomTkinter) & Render Pipeline
├── movie_review_v2.py          # V2 Engine: Scene Timeline Management & AI Prompting
├── voxcraft_license_client.py  # License Client Authentication Module
├── requirements.txt            # Python Dependencies
├── review_config.json          # Default Master Configuration
├── review_config.local.json    # User-defined Local Configuration
├── models/                     # Offline Faster-Whisper Speech Recognition Models
│   ├── base/                   # Base Whisper Model (config.json, model.bin, vocabulary)
│   ├── small/                  # Small Whisper Model
│   └── tiny/                   # Tiny Whisper Model
├── vendor/                     # Embedded Binary Tools & Offline Voice Engines
│   └── VoxCraft_TTS/           # Piper TTS Engine & Voice Packages
│       ├── piper/              # Piper C++ Executables, Libraries & Phonemizers
│       └── piper_vn/           # Vietnamese ONNX Voice Models (ngochuyen, maiphuong, etc.)
└── image/                      # Application Screenshots & Graphic Assets
    └── Screenshot 2026-07-31 005443.png

📦 Pre-packaged Downloads: Complete models and vendor standalone zip packages can be downloaded directly from VoxCraftRecap Releases.

---

🛠️ System Requirements & Installation

1. Prerequisites

• Python: 3.10 or higher
• FFmpeg & FFprobe: Must be installed and added to system PATH.

2. Installation

Clone the repository and install the required dependencies:

git clone https://github.com/K07VN/VoxCraftRecap.git
cd VoxCraftRecap
pip install -r requirements.txt

💡 Note: If you do not want to fetch large binary assets via Git LFS during cloning, you can download the full vendor and models zip archives directly from GitHub Releases and extract them into your project root directory.

CUDA Note: For CUDA GPU acceleration with Faster-Whisper, ensure CUDA toolkits and matching PyTorch packages are installed.

---

🚀 Getting Started

1. Launch the application:

   python ai_movie_review.py

2. Workflow:

   • Step 1: Select your source movie file (MP4, MKV, AVI, etc.).
   • Step 2: Provide movie context or extract subtitles automatically using Faster-Whisper.
   • Step 3: Configure your AI Model (Gemini Flash/Pro), choose a TTS voice actor, and adjust visual transformation options (Zoom, Watermark, Phantom Shift).
   • Step 4: Click "One-Click Recap & Render" to generate the final recap video.

---

💖 Support the Original Developer

If you find this software useful, please support the original developer by visiting their official website: 👉 Official Website: https://voxcraftone.com/

---

⚠️ Disclaimer

(The source code has been analyzed and reverse-engineered—if you encounter errors, fix them yourself.)

This codebase was decompiled and reconstructed for educational and technical research purposes. The maintainers take no responsibility for issues arising from commercial use.