{ }
</>
λ
#
Muhammad Asad Khan
Cloud · AI · Dev

LegalEaseAI

اردو قانونی معاون — Urdu Legal Assistant

LegalEaseAI Interface

The Problem & Solution

Millions sign contracts written in dense English legalese they cannot understand, leading to exploitation. LegalEaseAI bridges this gap by allowing users to upload any legal document and receive instant, plain-Urdu explanations of every clause, color-coded risk ratings, and a full Q&A chatbot that answers questions directly from the document itself.

Key Features

  • Risk Detection: Clauses auto-classified as High, Medium, or Safe.
  • Urdu Explanations: Plain, culturally appropriate Urdu.
  • RAG Chatbot: Ask anything in Urdu or English, grounded only in your document.
  • PDF Reports: Downloadable color-coded risk report.
  • Concurrent Processing: Heavy AI calls fire simultaneously using asyncio.

System Execution Previews

Urdu Retrieval-Augmented Generation (RAG) Chatbot

QA Chatbot Interface

Automated Contract Risk Audit Report Engine

Risk Assessment Report Outline

Technical Architecture

Architecture Diagram

The backend extracts text via pdfplumber and python-docx, parses it using a LangChain RecursiveCharacterTextSplitter, and indexes it with sentence-transformers into a FAISS vector space. It processes the text concurrently through Gemini/Groq APIs to generate Urdu insights and serve a RAG pipeline.

FastAPI LangChain FAISS RAG / Groq ES6+ Vanilla JS
View Live App ↗ GitHub Repo ↗