DeenPAL
An intelligent Islamic chatbot powered by RAG technology to deliver accurate Hadith-based answers
What is DeenPAL?
DeenPAL is a Retrieval-Augmented Generation (RAG) based chatbot designed to provide accurate and personalized Islamic information from trusted Hadith sources. Built on modern AI technologies, DeenPAL combines semantic search with large language models to deliver contextually relevant answers grounded in authentic Islamic texts from Sahih Bukhari and Sahih Muslim.Key Features
RAG Architecture
Leverages retrieval-augmented generation for accurate Hadith retrieval
MMR Retrieval
Uses Maximal Marginal Relevance for diverse and relevant results
Semantic Search
ChromaDB vector store with sentence-transformers embeddings
Interactive UI
Streamlit-based web interface with chat history
Quick Start
Get started with DeenPAL in just a few steps:Installation
Set up your environment and install dependencies
Quickstart Guide
Get DeenPAL running in minutes
Configuration
Configure API keys and settings
API Reference
Explore the complete API documentation
How It Works
DeenPAL processes Hadith collections through a sophisticated pipeline:- Data Loading - Extracts and processes Hadith PDFs with metadata preservation
- Semantic Splitting - Chunks text based on Hadith structure (chapters and book numbers)
- Vector Embedding - Generates embeddings using sentence-transformers
- Retrieval - Finds relevant Hadiths using MMR-based search
- Generation - LLM synthesizes answers with proper citations
DeenPAL uses the
all-MiniLM-L6-v2 model for embeddings and DeepSeek LLM via OpenRouter API for response generation.Explore the Documentation
Core Concepts
Learn about RAG, retrieval strategies, and prompt design
Guides
Step-by-step tutorials for setup and usage
Components
Deep dive into each system component