ai-agent-video-transcription 1.1.2

Multi-agent AI system for video transcription using OpenAI Whisper, ChromaDB, and AutoGen

🎬 Video Transcription Agent

Multi-agent AI system for video transcription using OpenAI Whisper, ChromaDB, and AutoGen

Transform your videos into accurate transcriptions with an intelligent multi-agent system that combines the power of OpenAI Whisper, persistent memory with ChromaDB, and advanced agent coordination.

✨ Key Features

🤖 Multi-Agent Architecture

• Intelligent Coordination: Specialized agents work together seamlessly
• Video Analysis: Extract metadata, duration, and quality information
• Audio Transcription: High-accuracy speech-to-text using OpenAI Whisper
• Text Processing: Automatic punctuation, formatting, and error correction
• Output Formatting: Multiple formats (SRT, VTT, TXT, JSON, CSV, DOCX)
• Quality Assurance: Built-in validation and quality assessment

🧠 Persistent Memory with ChromaDB

• Semantic Search: Find content using natural language queries
• Session History: Keep track of all transcriptions across sessions
• Metadata Storage: Rich information about videos and processing
• Trend Analysis: Usage statistics and pattern recognition

💬 Interactive CLI Experience

• Intuitive Commands: Simple, English-based command interface
• Auto-completion: Tab completion for commands and file paths
• Progress Indicators: Real-time progress bars and status updates
• Rich Interface: Beautiful colors and formatting with Rich library
• Error Handling: Graceful error management and recovery

🎯 Advanced Capabilities

• Multiple Video Formats: MP4, AVI, MOV, MKV, FLV, WMV, WebM, M4V
• Language Detection: Automatic or manual language selection
• Precise Timestamps: Word-level timing accuracy
• Batch Processing: Handle multiple videos simultaneously
• Quality Optimization: Automatic recommendations for better results

🚀 Quick Start

Installation

# Install from PyPI
pip install ai-agent-video-transcription

# Or install from source
git clone https://github.com/lopand-solutions/video-transcription-agent.git
cd video-transcription-agent
pip install -e .

Prerequisites

• Python 3.9+
• FFmpeg (for audio/video processing)

# Install FFmpeg
# macOS
brew install ffmpeg

# Ubuntu/Debian
sudo apt update && sudo apt install ffmpeg

# Windows (using Chocolatey)
choco install ffmpeg

Basic Usage

Interactive CLI (Recommended)

# Start the interactive CLI
vt-cli

# Or use the full command
video-transcription-agent

Available Commands:

🚀 Main Commands:

  • discovery      - Find videos in current directory
  • select <ID>   - Select a video from discovery results
  • transcribe    - Transcribe video file (always in Spanish)
  • analyze       - Analyze video metadata
  • search        - Search in saved transcriptions
  • history       - Show transcription history
  • status        - Show system status
  • help          - Show all commands
  • exit          - Exit the program

Command Line Usage

# Transcribe a video file
vt-cli transcribe path/to/video.mp4

# Analyze video metadata
vt-cli analyze path/to/video.mp4

# Search in transcriptions
vt-cli search "artificial intelligence"

# Show system status
vt-cli status

Programmatic Usage

import asyncio
from ai_agent_video_transcription import CoordinatorAgent

async def transcribe_video():
    # Configuration
    config = {
        'transcriber': {
            'model_size': 'base',  # tiny, base, small, medium, large
            'language': 'es'       # Force Spanish transcription
        },
        'formatter': {
            'output_directory': './output'
        },
        'output_formats': ['txt', 'srt']
    }
    
    # Create coordinator
    coordinator = CoordinatorAgent(config)
    
    # Execute transcription
    result = await coordinator.execute({
        'video_path': 'path/to/video.mp4',
        'output_name': 'my_transcription'
    })
    
    return result

# Run transcription
result = asyncio.run(transcribe_video())
print(f"Transcription completed: {result['final_outputs']}")

📖 Documentation

Configuration

Environment Variables

# Optional: Set OpenAI API key for advanced features
export OPENAI_API_KEY="your_api_key_here"

# Optional: Custom ChromaDB directory
export CHROMADB_PERSIST_DIRECTORY="./my_chroma_db"

Configuration File

Create a config.yaml file for advanced configuration:

transcriber:
  model_size: "base"  # tiny, base, small, medium, large, large-v2, large-v3
  language: "es"      # Force Spanish, or null for auto-detection
  temperature: 0.0

formatter:
  output_directory: "./output"
  supported_formats: ["txt", "srt", "vtt", "json"]

memory:
  persist_sessions: true
  session_timeout_minutes: 60

Supported Formats

Input Video Formats

• MP4, AVI, MOV, MKV, FLV, WMV, WebM, M4V

Output Formats

• TXT: Plain text transcription
• SRT: SubRip subtitle format
• VTT: WebVTT subtitle format
• JSON: Structured data with timestamps
• CSV: Comma-separated values
• DOCX: Microsoft Word document

Whisper Models

Model	Size	Speed	Accuracy	Use Case
tiny	39M	Fastest	Basic	Quick testing
base	74M	Fast	Good	Recommended
small	244M	Medium	Better	Balanced
medium	769M	Slow	High	Quality focus
large	1550M	Slowest	Highest	Best quality

🏗️ Architecture

Video Transcription Agent
├── 🤖 Multi-Agent System
│   ├── Coordinator Agent (orchestrates workflow)
│   ├── Analyzer Agent (video metadata extraction)
│   ├── Transcriber Agent (Whisper-based transcription)
│   ├── Processor Agent (text enhancement)
│   └── Formatter Agent (output generation)
├── 🧠 Memory System (ChromaDB)
│   ├── Persistent storage
│   ├── Semantic search
│   └── Session management
└── 💬 CLI Interface
    ├── Interactive commands
    ├── Progress indicators
    └── Error handling

🔧 Advanced Usage

Batch Processing

# Process multiple videos
videos = ['video1.mp4', 'video2.mp4', 'video3.mp4']
results = await coordinator.execute_batch(videos, 'batch_output')

Custom Memory Queries

# Search in stored transcriptions
memory = MemoryManager()
results = memory.search_transcriptions("machine learning", limit=10)

Quality Analysis

# Analyze video quality before transcription
analyzer = AnalyzerAgent()
metadata = await analyzer.execute({'video_path': 'video.mp4'})
print(f"Quality Score: {metadata['quality_score']}")

🐛 Troubleshooting

Common Issues

FFmpeg not found:

# Verify installation
ffmpeg -version

# Install if missing
brew install ffmpeg  # macOS
sudo apt install ffmpeg  # Ubuntu

Memory issues with large models:

• Use smaller models: tiny, base, small
• Process videos in smaller segments
• Increase system RAM or use GPU acceleration

Transcription in wrong language:

• Set language explicitly: language: "es" in config
• Check audio quality and clarity
• Try different Whisper models

Performance Optimization

• GPU Acceleration: Install CUDA-compatible PyTorch for faster processing
• Model Caching: Models are cached automatically after first use
• Batch Processing: Process multiple videos in sequence for efficiency

🤝 Contributing

We welcome contributions! Please see our Contributing Guide for details.

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

📝 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

• OpenAI Whisper - Speech recognition models
• ChromaDB - Vector database for semantic search
• AutoGen - Multi-agent framework
• Rich - Beautiful terminal interfaces
• FFmpeg - Multimedia processing

📞 Support

• 📧 Email: [e-mail]
• 🐛 Issues: GitHub Issues
• 📖 Documentation: Project Wiki

---

🎬 Transform your videos into accurate transcriptions with AI! 🚀

Made with ❤️ by Lopand Solutions