LLMFlow Search Agent is an intelligent research assistant that finds accurate answers to complex questions using advanced search strategies. The agent automatically refines queries, explores multiple information sources, and generates comprehensive reports with proper academic citations.

Key Features:

• Search Intent Analyzer optimizes queries for different search engines
• Multi-source search (DuckDuckGo, Wikipedia, web parsing)
• Intelligent planning system with LLM-guided strategy
• Academic-style report generation with numbered citations
• Selenium-based web scraping with caching
• Multiple LLM provider support (OpenAI, Anthropic, Gemini)

Example of a comprehensive report generated by LLMFlow Search Agent

• Python 3.9+
• Chrome/Chromium browser
• LLM API key (OpenAI, Anthropic, or Google Gemini)

git clone https://github.com/KazKozDev/LLMFlow-Search.git
cd LLMFlow-Search
pip install -r requirements.txt
echo "OPENAI_API_KEY=your_api_key_here" > .env

# Basic usage
python main.py

# With options
python main.py --output report.md --verbose --max-iterations 10

Configuration - Create config.json:

{
    "llm": {
        "provider": "openai",
        "model": "gpt-4o-mini",
        "temperature": 0.2,
        "max_tokens": 4096
    },
    "search": {
        "max_results": 5,
        "parse_top_results": 3,
        "use_selenium": true,
        "use_cache": true
    }
}

# Run tests
python -m pytest

# With coverage
python -m pytest --cov=core --cov=tools

# Verbose testing
python main.py --verbose

• Agent Core: Central coordinator managing information flow
• LLM Service: Unified interface for multiple LLM providers (OpenAI, Anthropic, Gemini)
• Search Intent Analyzer: Optimizes queries for different search engines
• Planning Module: Creates comprehensive search strategies
• Memory Module: Manages search context and results
• Report Generator: Synthesizes information into academic-style reports

• DuckDuckGo Search: Selenium-based scraping with anti-detection
• Wikipedia Tool: Multi-language Wikipedia integration
• Web Content Parser: Intelligent content extraction from web pages

graph TD
    A[User Query] --> B[Search Intent Analyzer]
    B --> C[Planning Module]
    C --> D[Agent Core]
    D --> E[Tools Module]
    E --> F[DuckDuckGo Search]
    E --> G[Wikipedia Search]
    E --> H[Web Content Parser]
    F --> I[Memory Module]
    G --> I
    H --> I
    I --> J[Report Generator]
    J --> K[Final Report]

• Never commit API keys to version control
• Use .env files for sensitive configuration
• All user inputs are validated and sanitized
• HTTPS-only connections for external requests

Reporting Vulnerabilities: Email security concerns to maintainers, do not create public issues.

• Use welcoming and inclusive language
• Be respectful of differing viewpoints
• Accept constructive criticism gracefully
• Focus on what is best for the community

# Fork repository and setup
python -m venv venv
source venv/bin/activate  # Windows: venv\Scripts\activate
pip install -r requirements.txt

• PEP8/flake8 compliance: 79 characters max, proper imports
• Documentation: All public functions must have docstrings
• Testing: Unit tests required for core functionality
• Error handling: Use specific exceptions, no bare except:

1. Create feature branch: git checkout -b feature/your-feature
2. Run linting: flake8 .
3. Run tests: python -m pytest
4. Submit pull request with clear commit messages

MIT License - see LICENSE file for details.

• GitHub Issues: Project Issues
• LinkedIn: Artem KK

⭐ If you find this project helpful, please give it a star