Loading page...
Sacred texts
Verses indexed
Embedding model
Retrieval method
Ancient wisdom explored through modern AI engineering — grounded, structured, and conversational.
Answers synthesized only from retrieved scripture — with inline citations, not speculation.
Primary and supporting teachings from Bhagavad Gita and Yoga Sutras with scripture badges.
BGE embeddings, BM25, cross-encoder reranking, and direct chapter/verse lookup.
Multi-turn memory and related follow-up questions guide deeper exploration.
Dynamic templates adapt to meaning, practice, comparison, and philosophical questions.
Docker, typed APIs, expanded evaluation metrics, and a polished product UI.
Full codebase available for review, learning, and contribution.
DHARMA (Divine Healing And Reflective Mindfulness Assistant) is a production-quality RAG application that helps you explore philosophical questions through verses from the Bhagavad Gita and Patanjali Yoga Sutras — with intent-aware answers, inline citations, and multi-turn memory.
Retrieval-Augmented Generation combines search with language models. Your question is routed by intent, verses are retrieved via metadata lookup or hybrid search with reranking, and an LLM synthesizes a structured answer with inline citations and follow-up questions.
The Bhagavad Gita offers practical wisdom on duty, devotion, and equanimity. The Yoga Sutras provide a systematic framework for consciousness and inner discipline — together forming a rich corpus for philosophical inquiry.
A retrieval-augmented experience with dynamic templates, cited sources, and follow-up discovery.
BGE embeddings, BM25 keyword search, cross-encoder reranking, and metadata-aware verse lookup.
Structured Markdown with inline citations, scripture badges, and collapsible verse cards.
Intent-aware prompts, multi-turn memory, related questions, FastAPI, and Next.js.
Three steps from question to cited, structured guidance.
Pose a question about dharma, yoga, a specific verse, or daily practice.
Metadata lookup, hybrid search, and reranking surface the best passages.
An intent-aware LLM synthesizes a clear answer with follow-up questions.
Start with a thoughtful inquiry — each answer is grounded in retrieved verses with related follow-ups.