mr-don88/Auto-generate-subtitles

🎤 Subtitle Generator

Automatically generate subtitles from audio/video files using OpenAI's Whisper model.

Features

• 🎙️ Speech-to-Text: Transcribe audio using Whisper
• 🌍 Auto Language Detection: Automatically detects source language
• 📝 Word-Level Timing: Shows timing for each individual word
• 📄 SRT Export: Download subtitles in standard SRT format
• 🎬 Video Support: Extract audio from video files (MP4, MOV, MKV, AVI)
• 📖 Story Mode: Merge short sentences into longer segments

How to Use

1. Upload an audio or video file

   • Audio: MP3, WAV, OGG, M4A, FLAC, AAC
   • Video: MP4, MOV, MKV, AVI

2. Select Whisper model size

   • tiny: Fastest, least accurate
   • base: Balanced (recommended)
   • small: More accurate, slower
   • medium: Most accurate, slowest

3. Choose processing mode

   • Podcast: Keep original sentence structure
   • Story: Merge short sentences (max 300 chars, 12 sec)

4. Click "Generate Subtitles" and wait for processing

5. Download the SRT file

Model Information

Model	Speed	Accuracy	Use Case
tiny	Fastest	Low	Quick previews
base	Fast	Good	General use (recommended)
small	Medium	Better	Podcasts, lectures
medium	Slow	Best	Professional transcription

Technical Details

• Uses faster-whisper for efficient CPU inference
• VAD filter removes silence/speech detection
• Word timestamps for precise alignment
• Auto language detection for multilingual support
• Smart number formatting (e.g., "1030" → "10:30" in time context)

Notes

• First transcription will download the model (~150MB for base model)
• Processing time depends on file length and model size
• Videos are processed by extracting audio first
• All processing happens on CPU (no GPU required)