A powerful video processing tool that generates transcripts with timestamps and keyframes with AI-generated descriptions.

• Automatic Speech Recognition: Extract transcripts with precise timestamps using OpenAI Whisper
• Keyframe Extraction: Extract frames at regular intervals from videos
• AI-Powered Image Descriptions: Generate detailed descriptions of keyframes using OpenAI's GPT-4 Vision
• Batch Processing: Process multiple videos at once
• Flexible Output Format: Generate structured output with timestamps

The tool generates output in this format:

[transcript:00:00:15] Hello, welcome to our video presentation
[keyframe:00:00:30] A professional speaker standing at a podium in a modern conference room with a large screen displaying charts
[transcript:00:00:45] Today we'll be discussing the latest trends in technology
[keyframe:00:01:00] Close-up shot of hands typing on a laptop keyboard with code visible on the screen

1. Install Python dependencies:

   pip install -r requirements.txt

2. Install FFmpeg (required for audio extraction):

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

3. Set up OpenAI API Key:

   cp env.example .env
   # Edit .env and add your OpenAI API key

python process_videos.py

python video_processor.py videos/your_video.mp4 -o outputs/output.txt

python video_processor.py videos/ -o outputs/

# Custom keyframe interval (every 60 seconds)
python video_processor.py video.mp4 -o output.txt -i 60

# Custom image description prompt
python video_processor.py video.mp4 -o output.txt -p "Describe what products are visible in this frame"

# Use API key directly
python video_processor.py video.mp4 -o output.txt --api-key sk-your-key-here

• -i, --interval: Keyframe extraction interval in seconds (default: 30)
• -p, --prompt: Custom prompt for AI image descriptions
• --api-key: OpenAI API key (alternatively set OPENAI_API_KEY environment variable)

• MP4, AVI, MOV, MKV, WMV, FLV, WebM

• Python 3.8+
• OpenAI API key with GPT-4 Vision access
• FFmpeg for audio processing
• Sufficient disk space for temporary audio files

• Whisper: Free (runs locally)
• GPT-4 Vision: ~$0.01-0.03 per image depending on detail level
• For a 10-minute video with 30-second keyframe intervals, expect ~20 API calls

1. "FFmpeg not found": Install FFmpeg and ensure it's in your PATH
2. "OpenAI API key missing": Set up your API key in .env file or use --api-key
3. "Out of memory": Use smaller Whisper model in code (change to 'tiny' or 'base')
4. "GPU errors": The tool works on CPU, no GPU required

• Use base Whisper model for good balance of speed/accuracy
• Increase keyframe interval for longer videos to reduce API costs
• Process videos in smaller batches if you have many files