Law, made legible
for everyone.
QanoonDaan is a free AI assistant that answers questions about Pakistani law in plain language — in English, Roman Urdu, or اردو.
Access to legal knowledge should not depend on who you can afford.
Pakistan has a 220 million strong population, and most people have never consulted a lawyer. Legal jargon, high fees, and language barriers keep ordinary citizens from understanding their own rights — even when those rights are being violated.
QanoonDaan was built to change that. By grounding AI responses in actual Pakistani statutes, it gives anyone with a phone the ability to ask a legal question and get a clear, cited, and honest answer — with a consistent reminder to consult a qualified lawyer for their specific situation.
You ask a question
Type or speak your question in any language — English, Roman Urdu, or Urdu script. The system detects your language automatically and handles the query accordingly.
Multilingual Input · Voice RecognitionRetrieval across Pakistani statutes
A hybrid retrieval engine combines BM25 keyword search with dense vector similarity to find the most relevant sections from a curated database of Pakistani laws — from the PPC to PECA to the Family Laws Ordinance.
BM25 · FAISS Vector Search · Section-Number BoostingGrounded answer generation
Retrieved sections are passed to a large language model (Llama 3.3-70B via Groq) with a hardened prompt that instructs the model to cite statutory sources, avoid hallucination, and always recommend professional legal advice for case-specific situations.
Llama 3.3-70B · Groq · Anti-Hallucination PromptingAnswer delivered — and also on WhatsApp
You get a clear, plain-language answer with a legal disclaimer. The same pipeline runs inside a WhatsApp bot via the Meta Cloud API, meaning QanoonDaan works on any phone without needing a browser.
FastAPI · WhatsApp Bot · Meta Cloud APIQanoonDaan's knowledge base currently covers the following Pakistani statutes. Answers are grounded only in these sources — if a question falls outside this scope, QanoonDaan will say so rather than guess.
Software Engineering graduate from the University of Sindh, Jamshoro, and alumnus of the SMIT AI & Data Science program. Built QanoonDaan from scratch — PDF ingestion, hybrid retrieval, multilingual inference, and WhatsApp integration — as a practical demonstration of what RAG-powered AI can do for underserved communities in Pakistan.
An Islamic finance platform with a RAG chatbot, Zakat calculator, halal stock screener, Mudarabah profit simulator, and a WhatsApp bot — all grounded in Shariah principles and deployed on Hugging Face Spaces.
An AI-powered resume screening API that scores candidate CVs against job descriptions using NLP. Built with FastAPI, spaCy, and Scikit-learn, containerised with Docker, and deployed on Hugging Face Spaces.