AI Video Agent v6 - Multi-Client Edition

An intelligent video generation and management system that creates, analyzes, and publishes AI-generated videos across multiple platforms via a Telegram bot interface.

🎯 Features

• Multi-Client Management: Handle multiple content creators and platforms from one system
• Intelligent Planning: LLM-based story and content planning
• Video Generation: AI-powered image and audio synthesis
• Smart Analysis: Video performance tracking and feedback integration
• Memory System: Learns from successful patterns and failures
• Daily Scheduling: Automated content generation and publishing
• Telegram Control: Command-based bot interface for easy management
• Reference Learning: Platform-specific reference video analysis

🏗️ Architecture

ai_video_agent/
├── agent.py              # Core orchestration engine
├── llm_planner.py        # Story and content planning
├── image_generator.py    # AI image synthesis
├── audio_generator.py    # Text-to-speech generation
├── video_assembler.py    # Video composition and export
├── video_analyzer.py     # Performance metrics and analysis
├── story_validator.py    # Content quality validation
├── telegram_bot.py       # Bot command interface
├── client_profile.py     # Multi-client management
├── memory_system.py      # Pattern learning system
├── daily_scheduler.py    # Automated scheduling
└── main.py               # Entry point

📋 Requirements

• Python 3.8+
• See requirements.txt for dependencies

🚀 Quick Start

1. Clone & Setup

git clone <repository-url>
cd AI_Video_Agent_v6
pip install -r requirements.txt

2. Configure

Create your client profiles in clients/ directory (YAML format):

name: "Your Client Name"
platform: "youtube"
content_style: "educational"
# See clients/example_fitness_coach.yaml for full template

3. Environment Setup

# Windows
setup_windows.bat

# macOS/Linux
python -m pip install -r requirements.txt

4. Run Options

Daily Scheduler (Automated generation):

python -m ai_video_agent.daily_scheduler
# or: run_daily.bat (Windows)

Telegram Bot (Interactive control):

python -m ai_video_agent.telegram_bot
# or: run_telegram.bat (Windows)

One-off Generation:

python -m ai_video_agent.main

⚙️ Configuration

config.yaml

Main configuration file with API endpoints, model settings, and platform defaults.

Client Profiles (clients/*.yaml)

Individual YAML files for each content creator:

• Content style and platform preferences
• Brand guidelines and tone
• Content frequency and scheduling
• Reference video patterns

📊 Memory & Learning

The system learns from:

• successful_patterns.json: High-performing content structures
• failed_patterns.json: What to avoid in future generations
• runs.json: Execution history and timing data
• feedback.json: User and audience feedback integration

🤖 Telegram Bot Commands

/start              - Initialize bot
/create [client]    - Generate new video
/status             - Current system status
/schedule [time]    - Set automated generation
/analyze [video_id] - Analyze video performance
/feedback [notes]   - Submit feedback
/help               - Show all commands

📁 Directory Structure

• ai_video_agent/ - Python package with core modules
• clients/ - Client configuration YAML files
• memory/ - JSON files for learning and history
• assets/ - Music tracks and sound effects
• reference_videos/ - Platform-specific reference content
• config.yaml - Main configuration

🔌 API Integration

The system integrates with:

• LLM APIs (for planning and validation)
• Image generation APIs
• Text-to-speech services
• Video hosting platforms
• Telegram Bot API

Configure API keys in config.yaml or environment variables.

📈 Performance Tracking

Videos are analyzed for:

• Engagement metrics
• Watch time patterns
• Audience retention
• Platform-specific performance
• Content pattern effectiveness

🛠️ Development

Adding New Platforms

1. Create platform adapter in client profile
2. Extend video_analyzer.py with metrics
3. Add reference videos in reference_videos/

Customizing Video Generation

1. Modify prompts in llm_planner.py
2. Adjust image/audio settings in generators
3. Update validation rules in story_validator.py

📝 Logging & Debugging

Check memory/runs.json for execution logs and timing. Errors are tracked in memory/failed_patterns.json with context.

🔐 Security Notes

• Store API keys in environment variables, not config.yaml
• Use Telegram bot tokens securely
• Keep client profiles out of version control if they contain sensitive data
• Add .gitignore entries for credentials

📄 License

[Specify your license]

🤝 Contributing

Contributions welcome! Please:

1. Fork the repository
2. Create feature branch
3. Test thoroughly
4. Submit pull request

📧 Support

For issues, questions, or feature requests, please open a GitHub issue.

🗺️ Roadmap

• Web dashboard for management
• Advanced analytics visualization
• Multi-language support
• Additional platform integrations
• Cloud deployment templates
• Advanced prompt engineering UI