📜 AI Legal Assistant – Misleading Clause Detector

An AI-powered web application that analyzes Terms & Conditions or legal documents (PDF/TXT) and detects misleading clauses. The system uses SentenceTransformer embeddings, FAISS similarity search, and an optional fine-tuned BERT classifier to identify risky or biased clauses.

🚀 Features

• 📂 Upload PDF or TXT contracts.

• 🔎 Detects misleading or biased clauses automatically.

• ✨ Highlights detected clauses directly inside the document.

• 📌 Provides a summary list of misleading clauses.

• ⚡ Uses FAISS + SentenceTransformer for fast similarity search.

• 🤖 Supports fine-tuned BERT classifier for better accuracy.

• 🌐 Web app built with Flask + HTML/CSS frontend.

🛠️ Tech Stack

• Backend: Python, Flask

• NLP Models: Hugging Face Transformers, Sentence-Transformers

• Vector Search: FAISS

• Frontend: HTML5, CSS3 (custom UI with gradient + animations)

• Libraries: PyPDF2, NLTK, NumPy, Pandas

⚙️ Installation

1️⃣ Clone the repository

git clone https://github.com/swalhasakeer/AI-Legal-Assistant-for-Detecting-Misleading-Clauses-in-Contracts.git
cd AI-Legal-Assistant-for-Detecting-Misleading-Clauses-in-Contracts

2️⃣ Create a virtual environment

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

3️⃣ Install dependencies

pip install -r requirements.txt

▶️ Usage

1. Run Flask app:

python app.py

2. Open in browser:

http://127.0.0.1:5000/

3. Upload a .pdf or .txt file.

4. View results:

   • 📄 Document text with highlighted misleading clauses.

   • 📌 Summary list of all detected clauses.

🎬 Demo

---

🔮 Future Improvements

✅ Mobile/Cloud deployment.

✅ Advanced explanation.

✅ Multi-language support.

✅ Enterprise integration.

✅ Clause recommentation engine

---