AuthentiText is a Streamlit-based web application designed to analyze and humanize text, transforming robotic or AI-generated content into natural, engaging, and authentic writing. It uses the OpenAI API to evaluate text authenticity with a "Human Score" and provides suggestions for improvement, while also offering a humanization feature to rewrite text in various styles.
- Text Analysis 🔍: Assesses text for human-like qualities with a strict scoring system (0-100% Human Score).
- Text Humanization ✍️: Rewrites text to sound natural and relatable in predefined or custom styles.
- Dual Theme Support 🎨: Offers light and dark themes with customizable CSS for a modern UI.
- Custom Styles 🎭: Allows users to define and save custom writing styles with examples.
- Actionable Feedback 💡: Provides detailed feedback and improvement suggestions based on analysis.
- Python 3.8+
- An OpenAI API key (stored in a
.envfile or Streamlit secrets)
- Clone the repository:
git clone <repository-url> cd authentiText
- Install dependencies:
pip install -r requirements.txt
- Create a
.envfile in the project root and add your OpenAI API key:AIMLAPI_KEY=your_api_key_here - Run the application:
streamlit run app.py
- Input Text ✏️: Paste your text into the text area on the main interface.
- Analyze Text 🔎: Click "Analyze Text" to get a Human Score and feedback.
- Humanize Text 🌱: Select a style (e.g., Conversational, Formal) and click "Humanize Text" to rewrite the text.
- Custom Styles 🎨: Use the expander to define and save a new style with a name and example.
- Re-Analyze 🔄: Review the humanized text and analyze it again for a new Human Score.
- API Key 🔑: Required and loaded from
.envor Streamlit secrets. - Model 🤖: Uses
gpt-3.5-turbofor analysis and humanization. - Themes 🌗: Switch between "Bright" and "Dark" via the sidebar settings.
app.py: Main application script containing UI and backend logic..env: Environment file for API key (not included in version control).requirements.txt: List of Python dependencies.
Feel free to fork the repository, submit issues, or create pull requests for enhancements or bug fixes.
[MIT License] - See the LICENSE file for details.
For questions or support, reach out via the project repository or email.
Made by VIK 🎉