TestMe - AI-Powered Quiz Generator

Transform any content into engaging quizzes with the power of AI

TestMe is a production-ready, full-stack web application that generates intelligent quizzes from multiple content sources using advanced AI. Built with modern technologies and designed for scalability, it supports YouTube videos, PDFs, images, and custom text to create engaging multiple-choice, true/false, and mathematical questions.

🎯 Perfect for: Educators, content creators, training organizations, and anyone looking to create interactive learning experiences.

🚀 Live Demo: https://test-me-beta.vercel.app

⚡ Quick Start

Want to get TestMe running quickly? Follow these essential steps:

# 1. Clone and install
git clone https://github.com/nosisky/testme.git
cd testme
npm install

# 2. Setup environment (copy and fill with your API keys)
cp env.example .env.local

# 3. Start development server
npm run dev

Essential environment variables to configure:

OPENAI_API_KEY - For AI quiz generation
GOOGLE_CLIENT_ID & GOOGLE_CLIENT_SECRET - For authentication
MONGODB_URI - For database storage
YOUTUBE_API_KEY - For YouTube video processing

📚 Detailed setup instructions are available in the Getting Started section below.

✨ Features

🎯 Core Functionality

Multi-Source Content: Generate quizzes from YouTube videos, PDFs (up to 5 pages), images, and custom text
AI-Powered Knowledge Gap Analysis: Automatically identifies learning gaps and key topics
Smart Question Generation: Creates multiple-choice, true/false, and mathematical questions
Adaptive Difficulty: Customizable difficulty levels (easy, medium, hard, expert)
LaTeX Math Support: Full mathematical equation rendering with MathJax

🎨 User Experience

Interactive Quiz Interface: Animated quiz-taking experience with real-time feedback
Responsive Design: Optimized for desktop, tablet, and mobile devices
Professional UI: Modern design with consistent theming
Quiz Sharing: Share quizzes via secure links
Performance Analytics: Detailed quiz statistics and user performance tracking

🔐 Security & Authentication

Google OAuth Integration: Secure authentication with Google accounts
User Management: Role-based access control (admin/user)
Data Privacy: Secure storage of user data and quiz results
Anonymous Quiz Taking: Option for users to take quizzes without registration

🤖 AI Integration

Multiple AI Providers: Support for OpenAI, Anthropic Claude, DeepSeek, and AWS Bedrock
Environment-Based Configuration: Secure API key management
Enhanced Large Quiz Generation: Dynamic token management for 15+ questions
Fallback Systems: Robust error handling with mock service support
Cost Optimization: Choose providers based on budget and performance needs

🚀 Demo

Visit our live demo: https://test-me-beta.vercel.app

Test Account: Use Google OAuth to create your account or try the demo content.

🛠 Technology Stack

Frontend

Framework: Next.js 15 with App Router
Language: TypeScript
Styling: SCSS Modules with CSS Variables
UI Components: Custom React components with responsive design
Math Rendering: MathJax for LaTeX equation support

Backend

Runtime: Node.js with Next.js API Routes
Database: MongoDB with Mongoose ODM
Authentication: NextAuth.js with Google OAuth
File Processing: PDF parsing, image analysis, YouTube transcript extraction

AI Integration

AI SDK: Vercel AI SDK for unified provider interface
Providers: OpenAI, Anthropic Claude, DeepSeek, AWS Bedrock
Vision: OpenAI GPT-4 Vision for image text extraction

Deployment & DevOps

Platform: Vercel (recommended) or Docker
Monitoring: Built-in analytics and error tracking
CI/CD: GitHub Actions ready

🚀 Getting Started

Prerequisites

Node.js: Version 18 or higher
npm or yarn: Package manager
MongoDB: Local instance or MongoDB Atlas
Git: Version control

Installation

Clone the repository

git clone https://github.com/nosisky/testme.git
cd testme

Install dependencies
```
npm install
```
Environment setup
```
cp env.example .env.local
```

Environment Configuration

Create a .env.local file in the root directory with the following configuration:

# =================================================================
# APPLICATION CONFIGURATION
# =================================================================
NEXTAUTH_URL=http://localhost:3000
NEXTAUTH_SECRET=your_nextauth_secret_here

# =================================================================
# AUTHENTICATION (Required)
# =================================================================
GOOGLE_CLIENT_ID=your_google_client_id
GOOGLE_CLIENT_SECRET=your_google_client_secret

# =================================================================
# DATABASE (Required)
# =================================================================
MONGODB_URI=mongodb://localhost:27017/testme
# Or for MongoDB Atlas:
# MONGODB_URI=mongodb+srv://username:password@cluster.mongodb.net/testme

# =================================================================
# EXTERNAL APIs (Required)
# =================================================================
YOUTUBE_API_KEY=your_youtube_api_key

# =================================================================
# AI PROVIDER CONFIGURATION
# =================================================================
# Choose your primary AI provider (openai, claude, deepseek, bedrock)
DEFAULT_AI_PROVIDER=openai

# Configure API keys for your chosen provider(s)
# OpenAI (Recommended - reliable, fast, good quality)
OPENAI_API_KEY=your_openai_api_key

# Anthropic Claude (Best for detailed explanations)
CLAUDE_API_KEY=your_claude_api_key

# DeepSeek (Most cost-effective option)
DEEPSEEK_API_KEY=your_deepseek_api_key

# AWS Bedrock (Enterprise-grade, requires AWS setup)
AWS_ACCESS_KEY_ID=your_aws_access_key_id
AWS_SECRET_ACCESS_KEY=your_aws_secret_access_key
AWS_REGION=us-east-1

# =================================================================
# DEVELOPMENT & TESTING
# =================================================================
# Set to 'true' to use mock AI service for testing (no API costs)
USE_MOCK_AI=false

# Set to 'development' for detailed logging
NODE_ENV=development

🤖 AI Provider Setup

OpenAI (Recommended)

Best for: General use, reliable performance, fastest responses

Models: GPT-4o-mini (default), GPT-4, GPT-3.5-turbo
Pricing: ~$0.0001 per 1K tokens (very affordable)
Setup: Get API key

DEFAULT_AI_PROVIDER=openai
OPENAI_API_KEY=sk-your-openai-key-here

Anthropic Claude

Best for: Detailed explanations, complex reasoning, educational content

Models: Claude-3-haiku (default), Claude-3-sonnet
Pricing: ~$0.00025 per 1K tokens
Setup: Get API key

DEFAULT_AI_PROVIDER=claude
CLAUDE_API_KEY=sk-ant-your-claude-key-here

DeepSeek

Best for: Cost-effective alternative, budget-conscious deployments

Models: deepseek-chat (default)
Pricing: ~$0.00014 per 1K tokens (most affordable)
Setup: Get API key

DEFAULT_AI_PROVIDER=deepseek
DEEPSEEK_API_KEY=sk-your-deepseek-key-here

AWS Bedrock

Best for: Enterprise use, enhanced security, compliance requirements

Models: Claude-3.7-Sonnet (via inference profile)
Pricing: Variable based on AWS pricing
Setup: Requires AWS account and Bedrock access

DEFAULT_AI_PROVIDER=bedrock
AWS_ACCESS_KEY_ID=your-aws-access-key
AWS_SECRET_ACCESS_KEY=your-aws-secret-key
AWS_REGION=us-east-1

Important: Bedrock uses inference profiles for newer Claude models. The system automatically uses us.anthropic.claude-3-7-sonnet-20250219-v1:0 which is the cross-region inference profile for Claude 3.7 Sonnet.

JSON Output: Since Bedrock Claude models don't have native JSON mode, TestMe uses clear delimiters (<JSON_START> and <JSON_END>) to ensure reliable JSON extraction from responses.

Supported regions: us-east-1, us-east-2, us-west-2 (automatically routed based on availability)

Provider Comparison

| Provider | Cost | Speed | Quality | Best Use Case | |----------|------|-------|---------|---------------| | OpenAI | 💰💰 | ⚡⚡⚡ | ⭐⭐⭐ | General purpo

TestMe

Install / Use

README