Fact Check Agent

An AI-powered fact-checking tool that verifies claims from URLs or text input using CrewAI and OpenAI's GPT-4.

Features

• URL fact-checking: Analyze articles and extract verifiable claims
• Text fact-checking: Process direct text input for claim verification
• Specialized handling for news sources (e.g., BBC)
• Multi-agent system using CrewAI for thorough fact verification
• Advanced web scraping with Firecrawl and fallback mechanisms
• DuckDuckGo search integration for claim verification
• Beautiful Streamlit interface

Setup

1. Clone the repository
2. Create a virtual environment:

python -m venv .venv
source .venv/bin/activate  # On Windows: .venv\Scripts\activate

3. Install dependencies:

pip install -r requirements.txt

4. Create a .env file with your API keys:

OPENAI_API_KEY=your_openai_api_key
FIRECRAWL_API_KEY=your_firecrawl_api_key  # Optional, will fallback to BeautifulSoup

⚠️ Security Note: Never commit your .env file or expose your API keys. The .gitignore file is configured to exclude sensitive files.

Usage

Run the Streamlit app:

python -m streamlit run app.py

The app will be available at http://localhost:8501

Components

• app.py: Main Streamlit interface
• fact_check_bot.py: Core fact-checking logic and CrewAI implementation
• Agents:
  • Content Extractor: Scrapes and cleans webpage content
  • Claim Identifier: Extracts verifiable claims
  • Fact Researcher: Gathers evidence using DuckDuckGo search
  • Claim Verifier: Analyzes claims against evidence
  • Credibility Summarizer: Provides final analysis

Input Methods

1. URL Input:

   • Paste a news article URL
   • Click "Check Credibility"
   • View extracted claims and verification results

2. Text Input:

   • Enter text directly
   • Click "Check Text"
   • View claims and verification results

Tests

Run the test suite with:

python -m pytest tests/

The tests cover:

• Basic functionality
• Credibility rating extraction
• Error handling for missing API keys

Requirements

Key dependencies (see requirements.txt for full list):

• Python 3.8+
• streamlit >= 1.31.0
• crewai >= 0.16.2
• openai >= 1.12.0
• langchain >= 0.1.0
• duckduckgo-search >= 7.5.2
• firecrawl >= 0.1.0 (optional)
• beautifulsoup4 >= 4.12.0
• python-dotenv >= 1.0.0

Security Best Practices

1. API Keys:

   • Store API keys in environment variables
   • Never hardcode sensitive information
   • Use .env file for local development
   • For production, use secure secret management

2. Input Validation:

   • All URLs are validated and sanitized
   • Text input is cleaned and validated
   • Error handling for malformed inputs

3. Dependencies:

   • Keep dependencies updated
   • Run pip install --upgrade periodically
   • Check for security vulnerabilities

4. Error Handling:

   • Proper error messages without exposing system details
   • Graceful fallback mechanisms
   • Comprehensive logging without sensitive information

Contributing

Contributions are welcome! Please read our Contributing Guide before submitting pull requests.

For a list of changes and upcoming features, see the Changelog.

License

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

Disclaimer

This tool is for educational and research purposes. Always verify critical information through multiple reliable sources. The accuracy of fact-checking results depends on the availability and reliability of online sources.