Automatically process meeting videos to extract key moments, topics, and takeaways. The tool generates a navigable summary webpage with video chapters.

• Extracts audio from video files
• Transcribes speech to text using insanely-fast-whisper
  • Uses Whisper large-v3 model by default for best accuracy
  • Optimized for both NVIDIA GPUs and Apple Silicon
  • Flash Attention 2 support for NVIDIA GPUs
  • Word-level timestamp support
• Analyzes content to identify:
  • Major topics with timestamps
  • Key moments within each topic
  • Actionable takeaways
• Adds chapter markers to the video
• Generates an interactive HTML summary
  • Clickable timestamps for video navigation
  • Organized by topics
  • Highlights key moments and takeaways

• Python 3.11+
• ffmpeg
• NVIDIA GPU or Apple Silicon Mac
• Conda (for environment management)

1. Install ffmpeg:

   # macOS
   brew install ffmpeg
   
   # Ubuntu/Debian
   sudo apt-get install ffmpeg
   
   # Windows
   # Download from https://ffmpeg.org/download.html

2. Set up the conda environment:

   # Create and activate the conda environment
   conda env create -f environment.yml
   conda activate video-auto-index

   This will automatically install all dependencies, including insanely-fast-whisper.

3. Set up your Anthropic API key:

   export ANTHROPIC_API_KEY='your-api-key'

Process a video file:

python -m src.main [video path] [--output-dir output] [--device-id DEVICE]

Device options:

• Default: Automatically uses MPS on Apple Silicon, CPU/CUDA on other systems
• --device-id 0: Force CPU/CUDA device
• --device-id mps: Force MPS device on Apple Silicon

For example:

# Use default device (auto-detected)
python -m src.main /path/to/video.mp4 --output-dir output

# Force CPU/CUDA device
python -m src.main /path/to/video.mp4 --output-dir output --device-id 0

# Force MPS device on Apple Silicon
python -m src.main /path/to/video.mp4 --output-dir output --device-id mps

This will:

1. Extract audio from the video
2. Transcribe the audio using insanely-fast-whisper
3. Analyze the meeting content for topics and key moments
4. Generate an HTML summary page

The output directory will contain:

• audio.wav: Extracted audio file
• audio_transcript.json: Transcribed speech with timestamps
• audio_subtitles.srt: Generated subtitles
• meeting_analysis.json: Extracted topics, moments, and takeaways

Final web output is stored in:

• <video_base_path>/<video_filename>_summary.html

The analysis JSON follows this structure:

[
  {
    "topic": "Topic description",
    "timestamp": "HH:MM:SS,mmm",
    "key_moments": [
      {
        "description": "Key moment description",
        "timestamp": "HH:MM:SS,mmm"
      }
    ],
    "takeaways": [
      "Actionable takeaway 1",
      "Actionable takeaway 2"
    ]
  }
]

The project is organized into modular components:

• video_processor.py: Handles video/audio operations
• transcriber.py: Speech-to-text conversion using insanely-fast-whisper
• key_moments.py: AI content analysis
• web_generator.py: HTML summary generation
• main.py: Pipeline orchestration

Each component can be run independently, allowing for flexible processing pipelines.

The project includes a comprehensive test suite covering all components:

Run all tests:

pytest

Run with coverage report:

pytest --cov=src tests/

Run specific test categories:

# Unit tests only
pytest -v -m "not integration"

# Integration tests only
pytest -v -m "integration"

• test_video_processor.py: Tests for video and audio processing
• test_transcriber.py: Tests for speech-to-text conversion
• test_key_moments.py: Tests for content analysis with API mocking
• test_web_generator.py: Tests for HTML generation
• test_main.py: Integration tests for the full pipeline

The test suite includes:

• Unit tests for each component
• Integration tests for the full pipeline
• API mocking for external services
• Fixture-based test data
• Error handling verification
• Edge case validation

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Add tests for new functionality
5. Ensure all tests pass
6. Submit a pull request