← All projects // #0013

Sanad - AI powered hadith discovery

TopcoatRustRAGAI-AssistantLLM

Sanad is an AI-powered hadith discovery platform that uses Retrieval-Augmented Generation (RAG) to help users explore the Sunnah through natural-language questions. Instead of relying solely on keyword searches, users can ask questions in their own words and discover relevant narrations from a collection of more than 44,000 hadiths across 15 collections.

The RAG pipeline retrieves relevant hadiths based on the user's query and uses them as the context for generating an answer. This approach keeps responses grounded in the underlying source material rather than relying solely on the language model's internal knowledge.

Key Features

  • Built an AI-powered RAG pipeline for hadith question answering.
  • Natural-language search for discovering relevant narrations.
  • Semantic retrieval of relevant hadiths based on user queries.
  • Source-grounded AI responses using retrieved hadiths as context.
  • Displays the original collection, book, hadith number, and grading for verification.
  • Arabic hadith text with English translations.
  • Supports 44,000+ narrations across 15 hadith collections.
  • Designed to reduce unsupported or hallucinated answers by grounding responses in retrieved sources.