2.4 KiB
2.4 KiB
Auto Clip 🎧✂️
Local, offline-friendly DJ-style album teaser builder using Python + FFmpeg.
This repository contains multiple versions of the Auto Clip workflow, each adding more intelligence and DJ-oriented features.
📁 Repository structure
Auto Clip/
├── V_1/
├── V_2/
├── V_3/
├── README.md ← this file
├── README_EN.md
└── README_DA.md
- V_1 — Minimal MVP / proof-of-concept
- V_2 — Track selection, best-of mode, DJ ordering
- V_3 — Harmonic mixing (Camelot), downbeat-ish snap, 2-pass loudness
📖 Documentation
Choose your preferred language:
- 🇬🇧 English documentation → README_EN.md
- 🇩🇰 Dansk dokumentation → README_DA.md
Each language README contains:
- Full explanation of V_1 / V_2 / V_3
- Install instructions
- Usage examples
- DJ / trance tuning tips
- Limitations and troubleshooting
🎶 What is Auto Clip?
Auto Clip scans a folder of audio tracks and automatically creates a short DJ-style teaser by:
- Finding musical highlights
- Snapping cuts to bar/phrase boundaries
- Optionally applying harmonic (Camelot) ordering
- Rendering a smooth teaser with crossfades
- Exporting WAV + MP3 plus a detailed JSON report
It is designed for:
- DJs
- Producers
- Album / EP promo teasers
- Old school trance & electronic music
Everything runs locally — no cloud, no uploads.
🤖 Optional: Ollama integration
If you run Ollama locally, Auto Clip can optionally generate:
- README content
- Promo text (SoMe / YouTube)
- Tracklists with timestamps
Recommended model:
llama3.1:8b-instruct-q4_0
Ollama is optional and not required for audio processing.
🚀 Quick start (example)
cd V_3
python dj_teaser_v3.py --tracks-dir ./tracks --select all --teaser 60 --bars 2 --preroll-bars 1 --avoid-intro 30 --harmonic
📜 License
Choose the license that fits your project (MIT is commonly used).
❤️ Credits
- FFmpeg
- librosa
- numpy
Built for DJs who like automation, control, and clean signal paths.