A Streamlit-based web app to detect whether a given piece of text is written by a human or generated by an AI language model like ChatGPT. The app uses a custom-trained Gradient Boosting model with linguistically-informed features for accurate and explainable predictions.
- 🔍 Real-time text classification (Human vs AI)
- 📊 Confidence scores with visual bar chart
- 🧠 Linguistic analysis using POS tagging and readability metrics
- 🌐 Streamlit web interface for easy interaction
- 🧪 Built on a hybrid feature set (TF-IDF + POS + TextStats)
- 📦 Model and vectorizer saved via
joblibfor fast inference
- Clone the repo:
git clone https://github.com/Nazeef1/AItextdetector.git
cd AItextdetector- Install dependencies:
pip install -r requirements.txt
python -m spacy download en_core_web_sm- Run the app:
streamlit run streamlit_app.pyOriginally trained on the AI_Human.csv. Now supports HC3 dataset from Hugging Face for future updates.
- Prediction: 🤖 AI-generated
- Confidence: AI =
0.93, Human =0.07
👉 Click on the image below to watch the demo!
Have suggestions or improvements?
You're welcome to fork the repo, open issues, or create pull requests. Contributions of all kinds are appreciated!