AI-Tutor: Personalized Learning Engine for Children with Autism Spectrum Disorder

AI-Tutor is a full-stack learning platform built to support children with Autism Spectrum Disorder through calm lessons, adaptive quizzes, emotion-aware pacing, parent progress tracking, and AI-assisted tutor conversations.

The goal is not to replace teachers, therapists, or caregivers. It is a supportive learning engine that helps each child practice at an appropriate pace while giving parents and educators clearer insight into progress, mood patterns, and topic readiness.

Objectives

• Deliver autism-friendly lessons with predictable structure, simple language, and clear visual feedback.
• Personalize quiz difficulty across easy, medium, and hard levels using child progress and emotion trends.
• Provide topic-based recommended YouTube practice links for colors, numbers, letters, shapes, emotions, and social skills.
• Track caregiver-reviewed video practice completion.
• Let teachers or therapists assign topic goals and target levels.
• Track learning sessions, emotions, quiz accuracy, stars, streaks, and rewards.
• Generate parent-friendly reports and weekly learning summaries.
• Offer an AI tutor chat experience that can answer questions, explain topics, and review progress.

Tools and Technologies Used

• React and Vite for the frontend
• Node.js and Express.js for the backend API
• MongoDB and Mongoose for database storage
• OpenAI SDK for AI tutor/report generation
• face-api.js for browser-based emotion detection support
• Postman for API testing
• VS Code for development
• Git and GitHub for version control
• npm for package management

Main Modules

Child Learning Engine

• Lessons for colors, numbers, letters, shapes, emotions, and social skills
• Adaptive quiz levels:
  • Easy: fewer choices and simpler questions
  • Medium: standard topic practice
  • Hard: more questions, stronger distractors, and reasoning prompts
• Recommended YouTube practice resource for each topic
• Video completion tracking after caregiver review
• Text-to-speech support for lesson items and quiz questions
• Star rewards after lesson completion

Emotion-Aware Support

• Optional camera-based emotion detection
• Support prompts when repeated negative emotion patterns are detected
• Calm reset activities such as breathing, movement breaks, and confidence-building tasks
• Difficulty adjustment when a learner appears frustrated, sad, or sleepy

Parent and Teacher Dashboard

• Child profile management
• Teacher/therapist goal assignment with topic and target level
• Emotion check-ins
• Accuracy, session, and practice-time analytics
• Daily quest and streak tracking
• Parent alerts and notifications
• Downloadable TXT, JSON, and PDF reports
• Weekly digest generation

AI Tutor Chat

• Child-aware chat context
• Progress-aware fallback responses
• AI-powered tutoring when the backend OpenAI key is configured
• Helpful prompts for lessons, feelings, reports, and next-step recommendations

Project Structure

AI-Tutor/
  backend/
    config/
    controllers/
    middleware/
    models/
    routes/
    services/
    tests/
    .env.example
    server.js
  public/
    models/
  src/
    api/
    components/
    pages/
    styles/
    utils/
  package.json
  vite.config.js
  README.md

Local Setup

1. Install Dependencies

PowerShell may block npm.ps1 on some Windows systems. If that happens, use npm.cmd.

npm.cmd install
npm.cmd --prefix backend install

2. Configure Backend Environment

Create backend/.env from the example file:

Copy-Item backend\.env.example backend\.env

Required/backend keys:

PORT=5000
MONGO_URI=mongodb://127.0.0.1:27017/ai-tutor
OPENAI_API_KEY=

OPENAI_API_KEY is optional for basic app exploration, but AI tutor/report features need it.

3. Start MongoDB

Start a local MongoDB server, or replace MONGO_URI with your MongoDB Atlas connection string.

4. Start Backend

npm.cmd --prefix backend start

Backend default URL:

http://localhost:5000

5. Start Frontend

Open a second terminal:

npm.cmd run dev

Frontend default URL:

http://localhost:3000

API Testing With Postman

Use Postman to test the backend endpoints while the server is running.

Common flows:

• Register/login a parent account
• Create a child profile
• Save a learning session
• Fetch progress analytics
• Generate reports
• Test AI tutor chat

Key endpoint groups:

• POST /api/auth/register
• POST /api/auth/login
• GET /api/auth/me
• GET /api/children
• POST /api/children
• GET /api/progress
• PATCH /api/progress
• POST /api/progress/lesson
• POST /api/session/save
• POST /api/report
• POST /api/report/weekly-digest
• POST /api/chat
• GET /api/notifications

Scripts

Frontend:

• npm.cmd run dev - start Vite development server
• npm.cmd run build - build production frontend
• npm.cmd run preview - preview production build
• npm.cmd run test:frontend - run frontend learning engine tests
• npm.cmd run test - run frontend and backend tests

Backend:

• npm.cmd --prefix backend start - start Express API server
• npm.cmd --prefix backend test - run backend tests

Recent Learning Upgrade

Each learning topic now includes:

• A recommended YouTube practice link
• A caregiver-reviewed completion marker
• Separate easy, medium, and hard question sets
• More questions as difficulty increases
• Distinct prompts at each level so repeated lessons feel less static

GitHub Upload Checklist

Before uploading or opening a pull request:

1. Confirm real .env files are not staged.
2. Run npm.cmd install and npm.cmd --prefix backend install if dependencies changed.
3. Run npm.cmd run test.
4. Run npm.cmd run build.
5. Review README.md, CONTRIBUTING.md, and SECURITY.md.
6. Commit only source, config, docs, lockfiles, and public assets. Do not commit node_modules/ or dist/.

Recommended Future Improvements

• Add a therapist/teacher mode for manually assigning goals.
• Add offline printable worksheets for each topic.
• Add stronger accessibility controls for contrast, reading speed, and sound sensitivity.
• Add automated frontend tests for lesson completion and quiz behavior.
• Replace YouTube search links with a curated, reviewed video table before production use.

Important Notes

• Do not commit real .env files or API keys.
• The camera is optional and should be used only with caregiver consent.
• YouTube links should be reviewed by a parent/teacher before a child watches independently.
• This app is an educational support tool, not a medical diagnostic system.