Aaroh — The Ascent of Understanding

Aaroh is an AI-powered learning assistant that helps students move beyond rote memorization toward true comprehension.
It breaks down complex academic text into simplified explanations, relatable analogies, and auto-generated quizzes, making learning engaging and accessible.

Built for the Viksit Bharat Challenge by Team Runtime Terror, Aaroh aims to democratize education by making complex knowledge simple and inclusive — for every learner, in every language.

---

Overview

The Problem

Students often struggle to grasp dense academic material, leading to surface-level understanding and dependence on memorization. Traditional educational resources are static and fail to adapt to individual learning needs.

The Solution

Aaroh provides a dynamic, AI-driven solution that allows a student to paste any complex academic text and instantly receive:

• A simplified version of the content.
• A real-world analogy to make the concept intuitive.
• An auto-generated quiz to test comprehension.
• Multilingual support for inclusivity.

By transforming passive reading into active understanding, Aaroh fosters curiosity, confidence, and critical thinking.

---

Tech Stack

• Frontend: Next.js (TypeScript, pnpm)
• Backend: Express.js (Node.js, pnpm)
• AI Service: Python (Flask + LLM integration)
• Containerization: Docker + Docker Compose
• CI/CD: GitHub Actions (build and push caching)

---

Folder Structure

.
├── app/         # Next.js frontend (UI for text input and quiz display)
│   └── Dockerfile
├── api/         # Express backend (routes + middleware)
│   └── Dockerfile
├── ai/          # Flask microservice (AI-powered text simplification, analogy, quiz generation)
│   ├── main.py
│   ├── requirements.txt
│   └── Dockerfile
├── docker-compose.yml
└── pnpm-workspace.yaml

---

Getting Started

Clone the repository

git clone https://github.com/upayanmazumder/aaroh.git
cd aaroh

Run with Docker

Build and start all services:

docker compose up --build

Then open:

• Frontend: http://localhost:3000
• API: http://localhost:4000
• AI Service: http://localhost:5000

---

How It Works

1. The frontend (Next.js) provides a simple interface for students to input text.

2. The backend (Express) receives the text and forwards it to the AI service.

3. The AI service (Flask) processes the input using language models to:

   • Simplify complex concepts.
   • Generate analogies for better recall.
   • Create short quizzes to test understanding.

4. The processed output is sent back through the backend and displayed in the UI.

---

Example Flow

1. User submits text to /api/learn.
2. The Express backend routes it to the Flask endpoint /process.
3. Flask AI module generates a simplified explanation, analogy, and quiz.
4. The processed result is displayed on the web app.

---

Docker Compose Overview

Service	Port	Description
app	3000	Next.js frontend
api	4000	Express backend
ai	5000	Flask AI microservice

Each container is modular, isolated, and deployable independently.

---

Development Notes

• Uses pnpm for dependency management.
• Python virtual environments are handled through Docker.
• GitHub Actions workflow supports build caching for faster CI/CD.
• Easily extensible with additional services (databases, vector stores, etc.).

---

Future Scope

• Integrate production-grade LLMs (OpenAI, HuggingFace, Ollama).
• Add multilingual voice input and output.
• Introduce user analytics and adaptive learning.
• Expand inclusivity through support for regional languages.
• Deploy to cloud providers (Fly.io, Railway, Render, etc.).

---

Vision

To make learning simple, personalized, and empowering — cultivating thinkers, not memorizers.

---

License

Licensed under the MIT License.