Sahayak AI

Empowering teachers in multi-grade classrooms through AI

Overview

Sahayak AI is a comprehensive AI-powered platform designed specifically for multi-grade classrooms in under-resourced Indian schools. It provides four specialized AI agents to assist teachers in planning, teaching, assessment, and administration.

Problem Statement

• 30%+ of India's government schools operate in Multi-Grade Multi-Level (MGML) model
• 9%+ are single-teacher schools
• Teacher shortages, lack of training, and absence of localized materials make effective teaching nearly impossible

Solution

An AI-powered multilingual mobile app that supports single teachers handling 3-5 grades simultaneously with:

• NCERT curriculum alignment
• Regional language support
• Automated administrative tasks
• Interactive content generation

Features Demo

Question Generation

The system generates contextually relevant questions based on NCERT curriculum:

Attendance Processing

Upload a class photo and get detailed attendance analysis:

Architecture

Quiz Generation Flow

The quiz generation pipeline uses NCERT context retrieval, Claude-based grounding, chunking, embeddings, and structured JSON extraction.

Manim Video Generation Flow

The animation workflow routes Claude Sonnet 4.5 through MCP tools to create, validate, and render Manim scenes.

Four AI Agents

1. शिक्षक मित्र (Shikshak Mitra) - Teaching Companion

• Weekly lesson scheduling and material preparation
• Smart blackboard layouts for multi-grade teaching
• Hyper-local content generation with cultural adaptation
• Manim animation creation for mathematical concepts
• Textbook photo scanning and digitization

2. परीक्षा गुरु (Pariksha Guru) - Exam/Quiz Agent

• Personalized quiz generation based on weak concepts
• NCERT-aligned exam paper creation
• OMR/OCR automated answer checking
• Multi-level difficulty adaptation
• Real-time performance analytics

3. प्रबंधन साथी (Prabandhan Saathi) - Admin Partner

• Photo-based face recognition attendance
• Automatic parent messaging (below 60% attendance)
• Performance dashboards and analytics
• Teaching improvement suggestions

4. सहायता चैट (Sahayata Chat) - Help Assistant

• Student curiosity question handling
• News-to-lesson content conversion
• Real-world connection generation
• Cultural context integration
• Personal teaching help chatbot

Technology Stack

Frontend:

• Flutter & Dart for cross-platform mobile app

Backend:

• Python with FastAPI
• Google Cloud Platform (Firestore, PostgreSQL, Object Store)
• Vertex AI for ML deployment

AI/ML:

• Gemini API (Large Language Model)
• Google Agent Development Kit
• PaliGemma (Vision-Language Model)
• Google ML Kit (Face Recognition)
• Manim (Mathematical Animation Engine)

Data:

• RAG System with NCERT knowledge retrieval
• Firestore Vector Database
• PostgreSQL for structured data

Getting Started

Prerequisites

• Python 3.12.9
• Google Cloud Platform account
• Google AI API key
• Required system dependencies for face recognition

Installation

1. Clone the repository

git clone https://github.com/divyanshkul/Sahayak.git
cd Sahayak

2. Install dependencies

pip install -r requirements.txt

3. Set up environment variables

# Create .env file
cp .env.example .env

# Configure your settings
GOOGLE_API_KEY=your_google_api_key
GCP_BUCKET_NAME=your_bucket_name
GCP_CREDENTIALS_PATH=path_to_service_account.json

4. Set up Google Cloud credentials

# Place your service account JSON file in the secrets/ directory
mkdir secrets/
# Copy your service account file to secrets/

5. Run the application

python main.py

The API will be available at http://localhost:4000

API Documentation

Once running, visit:

• Swagger UI: http://localhost:4000/docs
• ReDoc: http://localhost:4000/redoc

API Endpoints

Health Check

GET /api/v1/health/

Shikshak Mitra (Teaching Companion)

POST /api/v1/shikshak-mitra/generation-questions
Content-Type: application/json

{
  "question": "Generate questions related to triangles for class 5"
}

POST /api/v1/shikshak-mitra/generate-animation
Content-Type: application/json

{
  "prompt": "Create an animation showing how angles in a triangle add up to 180 degrees"
}

Prabandhan Saathi (Admin Partner)

POST /api/v1/prabhandhak/attendance/upload-photo
Content-Type: multipart/form-data

photo: [image file]
class_id: "class_123"

Key Features

Face Recognition Attendance

• Upload class photos for automatic attendance marking
• Advanced face recognition using trained models
• Generates attendance reports with visual feedback
• Automatic parent notifications for low attendance

Mathematical Animation Generation

• AI-powered Manim animation creation
• Mathematical concept visualization
• Automatic video generation and cloud storage
• Educational content localization

Question Generation with RAG

• NCERT curriculum-aligned question generation
• Context-aware retrieval from vector database
• In-context learning with SQL-backed question bank
• Multi-language support

Multilingual Support

• Hindi, Marathi, Kannada, Punjabi, and English
• Cultural context adaptation
• Regional example generation

Configuration

Environment Variables

Variable	Description	Required
GOOGLE_API_KEY	Google AI API key	Yes
GCP_BUCKET_NAME	Cloud Storage bucket	Yes
GCP_CREDENTIALS_PATH	Service account JSON path	Yes
MANIM_SERVER_PATH	Path to Manim MCP server	No
PYTHON_ENV_PATH	Python executable path	No

Face Recognition Setup

1. Create a train/ directory
2. Add student photos named as student_name.jpg
3. The system will automatically load and train face encodings

Google Cloud Agentic AI Day Hackathon

• Team: "The Fast and Fourier"
• Problem Statement: Empowering teachers in multi-grade classrooms

Made for teachers and students in multi-grade classrooms across India