Automatically transcribe and burn subtitles into videos using OpenAI Whisper — optimized for Indonesian language content. 100% free, no API key needed.
- 🎙️ Auto-transcription using stable-ts (enhanced Whisper)
- 🟡 Karaoke-style highlighting — active word turns yellow
- 📐 Dynamic subtitle sizing based on video resolution
- 🔥 Subtitles burned directly into output video via FFmpeg
- ⚡ Batch processing — drop multiple videos, get all outputs at once
| Input | Output |
|---|---|
| Raw video (no subtitles) | Video with auto-generated burned-in captions |
Buat struktur folder berikut di MyDrive:
MyDrive/
└── Whisper/
├── video/ ← taruh video input di sini
├── caption/ ← file .ass subtitle (auto-generated)
└── output/ ← video hasil akhir dengan subtitle
Folder
caption/danoutput/akan dibuat otomatis saat notebook dijalankan.
Taruh video yang ingin ditambahkan caption ke folder MyDrive/Whisper/video/.
Upload file GenerateCaption.ipynb ke Google Colab.
Runtime → Change runtime type → T4 GPU → Save
Runtime → Run all — notebook akan otomatis:
- Install dependencies (
stable-ts,ffmpeg) - Mount Google Drive
- Proses semua video di folder
video/ - Simpan hasil ke folder
output/
Video dengan subtitle sudah tersedia di MyDrive/Whisper/output/.
⚠️ Setelah selesai, matikan sesi Colab viaRuntime → Disconnect and delete runtimeagar kuota GPU gratis tidak terpakai sia-sia.
At the top of the main cell, you can adjust:
model = stable_whisper.load_model('medium') # tiny / base / small / medium / large
language = 'id' # language code, e.g. 'en', 'id', 'ja'| Model | Speed | Accuracy | VRAM |
|---|---|---|---|
tiny |
Fastest | Low | ~1 GB |
base |
Fast | Medium | ~1 GB |
small |
Moderate | Good | ~2 GB |
medium |
Slow | Great | ~5 GB |
large |
Slowest | Best | ~10 GB |
💡 100% Gratis — menggunakan Google Colab free tier (GPU T4) dan Google Drive. Tidak perlu bayar apapun.
- OpenAI Whisper — speech-to-text model
- stable-ts — word-level timestamp stabilization
- FFmpeg — video processing & subtitle burning
- Google Colab + Google Drive — cloud GPU runtime & storage
MIT