# How to Reuse This Pipeline

1. Copy `01_story/` as story template
2. Generate voiceover with edge-tts (zh-CN-XiaoxiaoNeural recommended)
3. Generate scene frames using Pillow (see logs/compose_video.py for reference)
4. Compose with FFmpeg: loop images -> concatenate -> add audio
5. Generate subtitles manually or via faster-whisper
6. Create review package
7. Deploy to Review Portal
8. Run Kimi audit