This is a Flask-based application that provides an intelligent, self-improving task execution system with dynamic agent selection (OpenAI, ElasticSearch, Claude/Anthropic) through a Universal Connector interface.
recursive-task-runner/
├── app.py # Main Flask application
├── universal_connector.py # Universal Connector class
├── agent_manager.py # Agent management and selection logic
├── task_classifier.py # Task classification system
├── pricing_calculator.py # Dynamic pricing logic
├── requirements.txt # Python dependencies
├── config.py # Configuration management
├── static/
│ ├── css/
│ │ └── styles.css # Frontend styling
│ ├── js/
│ │ └── main.js # Frontend JavaScript
│ └── images/
├── templates/
│ ├── index.html # Main interface
│ └── base.html # Base template
├── tests/
│ ├── test_app.py
│ ├── test_connector.py
│ └── test_agents.py
├── docs/
│ ├── API.md # API documentation
│ ├── SETUP.md # Setup instructions
│ └── ARCHITECTURE.md # System architecture
├── .env.example # Environment variables template
├── .gitignore # Git ignore file
├── README.md # Project documentation
└── docker-compose.yml # Docker configuration
- Clone repository
- Install dependencies:
pip install -r requirements.txt
- Copy
.env.example
to.env
and configure API keys - Run:
python app.py
- Open http://localhost:5000
- Dynamic AI agent selection (OpenAI, ElasticSearch, Claude)
- Universal Connector for multi-source data integration
- Recursive task execution and self-improvement
- Real-time pricing calculation
- Web-based interface
- Comprehensive logging and monitoring
- Create repository:
recursive-task-runner
in NeverShitty org - Import this gist structure
- Set up development environment
- Configure API keys and services
- Begin implementation phase
- All environment variables should be configured in Replit Secrets
- Use Replit's package manager for dependencies
- Flask app configured for Replit hosting
- Ready for immediate development continuation