📘 GurukulAI: Curriculum-Aligned Generative QA Platform for Indian School Students

GurukulAI is an open-source, AI-powered educational platform designed to assist students in the Indian school system with NCERT-aligned question answering. It features a fine-tuned LLaMA 3.1 8B model integrated into a Retrieval-Augmented Generation (RAG) pipeline, enabling students to ask questions, practice MCQs, and get chapter-specific answers.

---

📝 Dataset Summary

• Languages: English, Hindi
• Tags: education, ncert, rag, qa-dataset, bilingual, india
• Size: 18k+ Question–Answer–Context triples
• License: CC BY 4.0
• Access: HuggingFace Dataset

Note: NCERT textbook content is used under fair-use provisions for non-commercial, educational research purposes.

---

💻 Platform Features

• 🗣️ Multilingual QA Chat: Ask questions in English or Hindi
• 📸 Image-Based Doubt Resolution: OCR-based query from handwritten text
• 📄 Source Grounding: Chapter and context reference per response
• 📝 MCQ Generator: Topic-based quiz generation and feedback
• ✍️ Theory Answer Evaluation: LLM-suggested improvements
• 🔁 Follow-Up Learning: Suggests next questions for practice
• 🌐 Try it here: GurukulAI Web App

---

📂 Repository Structure

gurukul-ai/
├── backend/
│   ├── app.py
│   └── index_llama.py
│
└── frontend/
    ├── app.py
    ├── templates/
    │   └── index.html
    └── static/
        ├── css/
        │   └── style.css
        ├── js/
        │   └── script.js
        └── images/
            └── logo.png

---

📄 License

This dataset and codebase are distributed under the Creative Commons Attribution 4.0 International (CC BY 4.0) license.

---

📣 Citation

  title     = {GurukulAI: An Interactive AI-Driven Educational Platform for Indian Education System},
  author    = {Isha Narang and Sneh Gosai and Mayank Singh},
  year      = {2025},
  note      = {Available at https://lingo.iitgn.ac.in/gurukulai/}
}