AI Text Humanizer is a Streamlit-based web application that transforms AI-generated text into more natural, human-like writing using the phi3 language model and advanced text processing techniques.
# Install required dependencies
pip install -r requirements.txt
# Start Ollama with phi3 model
ollama run phi3
# Launch the application
streamlit run ai_humanizer.py
- Advanced text transformation using phi3 model
- Adjustable temperature control (0.1-1.0)
- Multiple humanization techniques:
- Smart typo insertion
- Natural punctuation variation
- Organic repetition
- Dynamic text formatting
- Real-time detection risk assessment
- Human-likeness scoring
- Comprehensive text metrics
- Actionable improvement suggestions
-
Text Humanization
- Open the Text Humanizer tab
- Input your AI-generated text
- Adjust settings in sidebar
- Click "Humanize Text"
-
Detection Check
- Navigate to AI Detection Check tab
- Paste text for analysis
- Review detailed metrics and scores
- Follow improvement suggestions
- phi3 model integration via Ollama API
- Advanced post-processing pipeline
- Real-time text analysis
- Metric calculation system
- Three-tab layout:
- Text Humanizer
- AI Detection Check
- About
- Interactive controls
- Real-time progress tracking
- Detailed statistics display
- Python 3.7+
- Ollama with phi3 model
- NLTK library
- Streamlit
- Internet connection for initial setup
- Local processing only
- No external API calls
- Data remains on your machine
- Open-source codebase
- Start with default settings
- Adjust temperature for desired variation
- Enable techniques selectively
- Use detection check for validation
- Iterate based on feedback
- Requires local Ollama setup
- Performance depends on hardware
- Results vary with input length
- Heuristic-based detection
This tool is intended for legitimate use cases only. Not for unethical purposes.
Pull requests are welcome. For major changes, please open an issue first to discuss proposed changes.
Made with ❤️ by an AI enthusiast