An AI-powered video editor that automatically creates cinematic videos with emotion detection, smart transitions, and dramatic effects.
- 🎬 AI Emotion Detection - Automatically detects emotions in videos using VideoMAE
- 🎨 Cinematic Filters - Applies emotion-matched filters (dramatic, warm, cool, vintage, etc.)
- ✨ Smart Transitions - Explosive openings, dramatic transitions, and emotional endings
- 💫 Visual Effects - Confetti, hearts, sparkles, light leaks, and more
- 🎵 Auto Music Selection - Matches background music to video emotion
- 📊 Quality Analysis - Selects best frames and segments automatically
- 🎥 1080p Output - High-quality Full HD videos
cd video_editor_pythonpython3 -m venv venv
source venv/bin/activate # On Windows: venv\Scripts\activatepip install -r requirements.txtNote: This will download large ML models (~500MB) on first run.
Put all your input videos in a folder (e.g., ./videos/):
videos/
├── video1.mp4
├── video2.mp4
└── video3.mp4
Put music files in a folder (default: ./music/):
music/
├── epic.mp3
├── calm.mp3
├── joyful.mp3
├── tense.mp3
└── neutral.mp3
Note: The script will automatically match music to detected emotions.
python main.py --input ./videosThis will:
- Process all videos in
./videos/ - Use music from
./music/(default) - Create
cinematic_output.mp4
python main.py --input ./videos --output my_video.mp4python main.py --input ./videos --music /path/to/music --output my_video.mp4- Loads videos from input folder
- Detects emotions using AI (epic, calm, joyful, tense, neutral)
- Finds best segments using scene detection and quality analysis
- Applies filters matched to emotions
- Adds transitions - explosive openings, dramatic cuts, emotional endings
- Adds effects - sparkles, hearts, confetti, light leaks
- Selects music based on dominant emotion
- Renders a polished 1080p cinematic video
- Resolution: 1080p Full HD
- Duration: ~2.5 seconds per video segment
- Format: MP4 (H.264)
- Effects: Dramatic transitions, visual effects, emotion-matched filters
- Python 3.8+
- FFmpeg (usually installed with moviepy)
- ~2GB disk space for models and processing
- GPU optional (works on CPU, but slower)