🍎 Memories Creator

Professional video slideshow creation with Apple-style transitions and AI-generated music, optimized for Apple Silicon.

Transform your photo collections into stunning cinematic videos with AI-powered music generation, professional transitions, and automatic photo enhancement—all from the command line or Python API.

---

🌟 Features

The only open-source slideshow tool with built-in AI music generation.

🎬 Professional Video Creation

• 9 Transition Types: Ken Burns, zoom in/out, pan, fade in/out, cross-fade, dissolve, static
• Smart Photo Enhancement: Automatic enhancement for photos taken before 2020
• Hardware Acceleration: Native MPS support for Apple Silicon (M1/M2/M3/M4)
• Multiple Codecs: H.264, H.265 (HEVC) with VideoToolbox hardware encoding

🎵 AI Music Generation

• Text-to-Music: Generate custom soundtracks from natural language descriptions
• Multiple Models: Small (fast), Medium (balanced), Large (best quality), Melody (melody-aware)
• Seamless Looping: Intelligent audio extension for any video length
• Professional Mixing: Automatic volume balancing, fade in/out, crossfades

⚡ Performance

• GPU Acceleration: PyTorch MPS backend for transitions and effects
• Multi-threaded: Parallel image processing
• Memory Efficient: Handles 100+ photos with 16GB RAM

---

📦 Installation

From PyPI (Recommended)

pip install memories-creator

From Source

git clone https://github.com/kubataba/memories-creator.git
cd memories-creator
pip install -r requirements.txt
pip install -e .

Additional Requirements

FFmpeg (required):

# macOS
brew install ffmpeg

# Ubuntu/Debian
sudo apt install ffmpeg

---

📁 Итоговая структура проекта для PyPI

memories-creator/
├── memories_creator/        
│   ├── __init__.py
│   ├── cli.py
│   ├── main.py
│   ├── config.py
│   ├── music_generator.py
│   ├── video_processor.py
│   ├── transitions.py
│   ├── slide_editor.py
│   └── media_utils.py
│
├── README.md                    
├── LICENSE                       
├── setup.py                     
├── pyproject.toml 
├── config.json             
├── requirements.txt               
├── MANIFEST.in                   
└── .gitignore

---

🚀 Quick Start

1. Create Your First Video.

# Create configuration
memories config create

# Add your photos to ./photos directory
mkdir photos
cp ~/Pictures/*.jpg photos/

# Generate video
memories create

Output: output/memories_video.mp4

2. With AI Music.

Edit config.json:

{
  "art_dir": "./photos",
  "output_dir": "./output",
  "generate_music": true,
  "music_prompt": "uplifting piano melody, emotional strings, cinematic",
  "music_model": "small"
}

memories create

3. Command Line Options

# Custom paths
memories create --input ./vacation --output ./videos

# Custom music
memories create --music ./my-song.mp3

# Different transition
memories create --transition ken_burns

# Generate music only
memories music "epic orchestral" --duration 120

# List slides
memories list

# Fix rotations
memories fix 8:180 12:90

# Replace audio
memories audio --file new-music.mp3

---

📖 Python API

Basic Usage

from memories_creator import MemoriesCreator, Config

# Load configuration
config = Config.from_file('config.json')

# Create video
creator = MemoriesCreator(config)
creator.create_video()

Generate Music

from memories_creator import MusicGenerator

# Create generator
generator = MusicGenerator(model_name='small')

# Generate soundtrack
music_path = generator.generate(
    prompt="peaceful piano with gentle strings",
    duration=120,
    output_path="output.wav"
)

Batch Processing

from memories_creator import MemoriesCreator, Config

projects = ['vacation', 'wedding', 'family']

for project in projects:
    config = Config()
    config.set('art_dir', f'./{project}')
    config.set('output_dir', f'./output/{project}')
    
    creator = MemoriesCreator(config)
    creator.create_video()

---

🎛️ Configuration

Example config.json

{
  "art_dir": "./photos",
  "output_dir": "./output",
  
  "music_prompt": "nostalgic piano, gentle strings",
  "generate_music": true,
  "music_model": "small",
  "music_seamless": true,
  "music_volume": 0.3,
  
  "max_video_size": 1920,
  "fps": 30,
  "seconds_per_photo": 3,
  
  "transition_type": "ken_burns",
  "transition_intensity": 0.1,
  
  "enhance_old_photos": true,
  "year_threshold": 2020,
  
  "video_codec": "hevc_videotoolbox",
  "video_quality": 23
}

Transition Types

• ken_burns - Slow zoom + pan (classic)
• zoom_in - Smooth zoom in
• zoom_out - Smooth zoom out
• pan - Horizontal/vertical movement
• fade_in - Fade from black
• fade_out - Fade to black
• fade_cross - Crossfade between images
• dissolve - Smooth blend
• static - No movement (fast)

Music Models

• small - Fast, ~1.5GB (recommended)
• medium - Balanced, ~3.3GB
• large - Best quality, ~6.7GB
• melody - Melody conditioning

---

💻 System Requirements

Minimum:

• Python 3.9+
• 8GB RAM
• macOS 12.3+ or Linux (Ubuntu 20.04+)
• FFmpeg

Recommended:

• Python 3.10+
• 16GB+ RAM
• Apple Silicon (M1/M2/M3/M4) Mac
• macOS 13+
• 20GB disk space (for AI models)

---

🐛 Troubleshooting

MPS not available

python -c "import torch; print(torch.backends.mps.is_available())"

Update macOS to 12.3+ and PyTorch:
pip install --upgrade torch

Out of memory

Reduce max_video_size to 1280 and use transition_type: "static"

Music generation fails

Check disk space (models are 1.5-6GB): df -h

FFmpeg not found

# macOS: brew install ffmpeg
# Linux: sudo apt install ffmpeg

---

📄 License

CC-BY-NC-4.0 (Creative Commons Attribution-NonCommercial 4.0 International)

This project is licensed for non-commercial use only.

You are free to:

• ✅ Use for personal videos and memories
• ✅ Modify and adapt the code
• ✅ Share with attribution

You must:

• 📝 Give appropriate credit
• 🔗 Provide a link to the license
• ⚠️ Not use for commercial purposes

Attribution:

Memories Creator by Eduard Emkuzhev
License: CC-BY-NC-4.0
https://github.com/kubataba/memories-creator

Full license: https://creativecommons.org/licenses/by-nc/4.0/

---

🙏 Acknowledgments

• PyTorch - Deep learning framework
• MusicGen - AI music generation (Meta AI)
• FFmpeg - Audio/video processing
• OpenCV - Video encoding

---

📞 Support

• 🐛 Report Issues
• 💬 Discussions
• 📖 Documentation

---

🤝 Contributing

Contributions welcome! This project is licensed under CC-BY-NC-4.0.

git clone https://github.com/kubataba/memories-creator.git
cd memories-creator
pip install -r requirements.txt
pip install -e .

---

Made with ❤️ to keep your memories alive