Kanoon AI

Kanoon AI is an intelligent legal assistant on the Indian Law. It leverages advanced AI models and a robust backend to provide users with accurate and comprehensive answers to queries related to the Basic Indian Law.

Table of Contents

• Introduction
• Features
• Tech Stack
• Architecture

Introduction

Kanoon AI aims to simplify legal understanding for the common man by providing a user-friendly platform to query and receive detailed responses about the Indian Constitution. It combines the power of natural language processing with the richness of the Indian legal system.

Features

• Query the Indian Law: Ask questions and receive detailed responses.
• Streamlined Responses: Uses a Retrieval-Augmented Generation (RAG) model to ensure accurate and relevant answers.
• User-Friendly Interface: A clean and intuitive interface built with Next.js and Shadcn for styling.
• Persisted Storage: Utilizes Chroma DB for efficient storage and retrieval of legal documents.

Tech Stack

• Frontend: Next.js, React, Shadcn
• Backend: FastAPI
• Database: Chroma DB
• Machine Learning Models: Google Generative AI, Sentence Transformers

Architecture

The architecture of Kanoon AI is designed to be modular and scalable, adhering to SOLID principles.

1. Frontend: Built using Next.js with a focus on client-side rendering , streaming responses and static site generation.
2. Backend: FastAPI serves as the backend framework, providing RESTful API endpoints.
3. Database: Chroma DB is used for storing and retrieving embeddings of the documents.
4. Machine Learning: Utilizes Google Generative AI for generating responses based on retrieved context.
5. Document Processing: Documents are loaded and split using PyPDFLoader and text splitters, with embeddings generated using Sentence Transformers.

Demo

kanoon-ai-demo-linkedin.mp4