Covenant AI

Covenant AI is a smart lease contract analysis system that leverages Agentic RAG for information retrieval. The RAG architecture is built without the use of external libraries/frameworks for scalability.

For a comprehensive overview, please visit the offical Covenant AI Documentation

Demo Video

Click here to see the demo

Sample Dataset for Hackathon

The project root has a few sample lease agreements you can use to test the project and its funcationalities.

Features

Frontend (HTML/CSS/JavaScript)

• Contract Upload: Users can upload lease contracts in PDF or DOCX format.
• Summary & Insights: Provides an AI-generated summary of key terms.
• Highlight PDF: Highlights the important clauses and obligations in the pdf.
• Interactive Chatbot: Allows users to ask AI-driven questions about the contract.

Backend (Flask Framework)

• Document Processing: Extracts and parses text from uploaded documents using OCR.
• NLP Analysis: Uses AI models to analyze legal language.
• Database Management: Stores analyzed contracts in the supabase.
• Dynamic ground rules: Update rules and regulations that must be present in agreement.

AI Components

1. Summarization: Summarize the uploaded agreements with LLM.
2. Compare agreements: Compare 2 agreements and see which one is better.
3. ChatBot: RAG based chatbot that helps understand jargons.
4. Multiple contracts: Upload multiple PDFs and can chat with any of them.

Technologies Used

• Frontend: HTML, CSS, JavaScript
• Backend: Flask Framework, SupaBase
• AI & NLP: SpaCy, Transformers (BERT/GPT), OCR for text extraction
• Cloud & APIs: AWS S3 for file storage, FastAPI for AI models

📜 Get Started

Prerequisites

• Python 3.x
• SupaBase
• Flask framework
• NLTK, Sentence Transformers
• HTML, CSS, JS

Installing requirements

pip install -r requirements.txt

Installing Tesseract

Below are platform-specific instructions:

Windows

1. Download and install from: https://github.com/UB-Mannheim/tesseract/wiki
2. Ensure the tesseract.exe location is added to your PATH.

macOS

1. Install Homebrew if needed (https://brew.sh).
2. Run: brew install tesseract.

Linux (Ubuntu/Debian)

1. Update your package list: sudo apt-get update.
2. Install Tesseract: sudo apt-get install tesseract-ocr.

After installation, verify with:
tesseract --version

Running project

root/src: python app.py

License

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

Contact

For inquiries, contact u2022378@giki.edu.pk