A powerful, multi-language voice-to-text transcription tool that automatically detects languages and translates audio content to English. Built with OpenAI's Whisper API and GPT-4 for high-accuracy transcription and translation.

• 🌍 Multi-language Support: Automatically detects and transcribes audio in any language
• 🇺🇸 English Translation: Translates all detected languages to English
• 🎵 Multiple Audio Formats: Supports both MP3 and OPUS audio files
• 🔄 Automatic Conversion: Seamlessly converts OPUS files to MP3 before processing
• 📊 Detailed Analysis: Provides language detection, timestamps, and segment analysis
• 💾 JSON Export: Saves all transcription results in structured JSON format
• 🔒 Secure: Uses environment variables for API key management

• Python 3.11 or higher
• FFmpeg (for audio conversion)
• OpenAI API key

1. Clone or download the project files

2. Install Python dependencies:

   pip install -r requirements.txt

3. Install FFmpeg (if not already installed):

   • macOS: brew install ffmpeg
   • Ubuntu/Debian: sudo apt install ffmpeg
   • Windows: Download from ffmpeg.org

4. Set up your API key: Create a .env file in the project root:

   OPENAI_API_KEY=your_openai_api_key_here

voice-to-text-transcriber/
├── main.py                          # Main transcription script
├── opus_to_mp3_converter.py        # OPUS to MP3 converter utility
├── requirements.txt                 # Python dependencies
├── .env                            # Environment variables (create this)
├── README.md                       # This file
└── audio_files/                    # Your audio files go here
    ├── hindi-demo-voice.opus
    ├── japanese-text.opus
    └── other-audio-files.mp3

1. Place your audio file in the project directory
2. Update the filename in main.py (line ~200):

   audio_file = "your-audio-file.opus"  # or .mp3

3. Run the script:

   python3 main.py

• Input: .opus, .mp3
• Output: Automatically converts OPUS to MP3, then processes for transcription

🎤 Multi-Language Voice to English Transcription
==================================================
🎵 Detected .opus file: hindi-demo-voice.opus
🔄 Converting hindi-demo-voice.opus to MP3 format...
✅ Conversion successful: hindi-demo-voice.mp3

🔄 Attempting multilingual analysis...
🎵 Analyzing multilingual audio: hindi-demo-voice.mp3
🔄 Starting advanced language analysis...
✅ Advanced transcription completed!
🔍 Analyzing language segments...
🌍 Detected languages: ['hindi']

==================================================
📊 TRANSCRIPTION RESULTS
==================================================
🌍 Languages detected: hindi

📝 English Translations:

[HINDI] → English:
Hello, how are you today? I hope you're doing well.

💾 Results saved to: transcription_results.json

Create a .env file with:

OPENAI_API_KEY=your_openai_api_key_here

The script uses:

• Whisper Model: whisper-1 for transcription
• GPT Model: gpt-4o-mini for translation
• Temperature: 0.0 for transcription, 0.1 for translation

Contains structured data:

{
  "detected_languages": ["hindi"],
  "language_segments": {
    "hindi": [
      {
        "start": 0.0,
        "end": 3.5,
        "text": "नमस्ते, कैसे हो आप?"
      }
    ]
  },
  "english_translations": {
    "hindi": "Hello, how are you?"
  },
  "full_transcript": "नमस्ते, कैसे हो आप?"
}

The opus_to_mp3_converter.py utility provides:

• High-quality MP3 conversion
• Batch processing capabilities
• Configurable quality settings
• Metadata preservation

Usage:

# Single file conversion
python3 opus_to_mp3_converter.py input.opus

# Batch conversion
python3 opus_to_mp3_converter.py input_directory --batch

# Quality options
python3 opus_to_mp3_converter.py input.opus -q high    # Best quality
python3 opus_to_mp3_converter.py input.opus -q medium  # Good quality
python3 opus_to_mp3_converter.py input.opus -q low     # Standard quality

The main script automatically:

• Detects language changes in audio
• Segments audio by language
• Translates each segment to English
• Provides comprehensive analysis

1. "FFmpeg not found":

   • Install FFmpeg: brew install ffmpeg (macOS) or sudo apt install ffmpeg (Ubuntu)

2. "API key not found":

   • Check your .env file exists and contains OPENAI_API_KEY=your_key

3. Audio file not found:

   • Verify the filename in main.py matches your actual audio file

4. Conversion failed:

   • Ensure FFmpeg is properly installed
   • Check audio file format and integrity

• ❌ Audio file not found: Check filename and path
• ❌ Conversion failed: Verify FFmpeg installation
• ❌ API key not found: Check .env file
• ❌ Transcription failed: Check API key validity and audio quality

• Audio Quality: Higher quality audio = better transcription accuracy
• File Size: Larger files take longer to process
• Language Complexity: Some languages may require more processing time
• API Limits: Be mindful of OpenAI API rate limits and costs

• Never commit your .env file to version control
• Keep your API key private and secure
• Use environment variables for all sensitive configuration

This project is for educational and personal use. Please respect OpenAI's terms of service and API usage policies.

Feel free to:

• Report bugs and issues
• Suggest new features
• Improve documentation
• Optimize performance

For issues or questions:

1. Check the troubleshooting section
2. Verify your setup and configuration
3. Check OpenAI API status and limits

Happy Transcribing! 🎵✨