KellyKiprop/Legal-AI-assistant

A production-ready Retrieval-Augmented Generation (RAG) system that explains the Constitution of Kenya in plain English. # 🇰🇪 Kenya Constitution AI Chatbot A production-ready **Retrieval-Augmented Generation (RAG)** system that allows users to ask questions about the Constitution of Kenya and receive **clear, plain-English explanations** — understandable by anyone, not just legal experts. --- ## 🚀 Overview This project transforms the Constitution of Kenya from a dense legal document into an **interactive AI assistant**. Users can: * Ask legal questions in natural language * Get simplified explanations * See the exact constitutional articles backing the answer --- ## 🧠 Architecture ``` Frontend (Lovable UI) ↓ FastAPI Backend (api.py) ↓ FAISS Vector Search ↓ Relevant Articles Retrieved ↓ Groq LLM (LLM reasoning + simplification) ↓ Structured Plain-English Response ``` --- ## ✨ Features * 📄 **Article-level retrieval** (high accuracy vs chunk-based noise) * ⚡ **Fast FAISS similarity search** * 🧠 **LLM-powered explanations (Groq)** * 🗣️ **Plain English output (non-legal language)** * 🔍 **Cited constitutional sources** * 🔌 **API-first design (ready for any frontend)** --- ## 📂 Project Structure ``` . ├── api.py # FastAPI backend ├── rag_answer_engine_groq.py # Core RAG pipeline ├── extract_from_pdf.py # Extract raw text from PDF ├── chunk_by_article.py # Split into articles ├── clean_articles.py # Clean & normalize text ├── build_index.py # Create FAISS index ├── constitution.index # Vector index ├── constitution_articles_clean.jsonl ├── requirements.txt ├── README.md └── .gitignore ``` --- ## ⚙️ Setup ### 1. Clone repo ```bash git clone github.com cd kenya-constitution-rag ``` ### 2. Create environment ```bash python -m venv venv source venv/bin/activate ``` ### 3. Install dependencies ```bash pip install -r requirements.txt ``` ### 4. Set API key Create a `.env` file: ``` GROQ_API_KEY=your_api_key_here ``` --- ## ▶️ …