A-Guy: AI-Powered Educational Platform

A modern, AI-driven learning platform built with Payload CMS, Next.js, and MongoDB Atlas. AGuy is designed as an "AI Operating System" for education — combining a learning management system, intelligent chat with memory, content management, and multi-tenant infrastructure.

Overview

AGuy is not a typical website or LMS. It's a unified platform that integrates:

• Course Management: Hierarchical content structure (Courses → Chapters → Lessons → Exercises)
• AI-Powered Chat: Smart tutoring with context awareness and long-term memory
• PDF Processing: Extract exercises from PDF documents using Vision AI
• Multi-Tenant: Support for multiple organizations with isolated data
• Admin Panel: Full-featured CMS with custom components

📖 Read the full introduction — Learn about A-Guy's unique approach, advantages, and technical architecture.

Tech Stack

Layer	Technology
CMS & Data	Payload CMS 3.73
Frontend	Next.js 15 (App Router)
Database	MongoDB Atlas (with Vector Search)
Styling	Tailwind CSS + shadcn/ui
AI	Google Gemini, OpenAI
Deployment	Vercel

Quick Start

Prerequisites

• Node.js 18+
• pnpm 9+
• MongoDB Atlas account (with Vector Search enabled)
• Google Gemini API key (for AI features)

Setup

# Clone the repository
git clone https://github.com/A-Guy-educ/A-Guy.git
cd A-Guy

# Install dependencies
pnpm install

# Copy environment variables
cp .env.example .env

# Start MongoDB (local) or ensure Atlas is configured
docker-compose up -d

# Generate types and import map
pnpm generate:types
pnpm generate:importmap

# Start development server
pnpm dev

Open http://localhost:3000 to access the application.

Admin Access

The admin panel is available at /admin. Create your first admin user during the initial setup.

Project Structure

src/
├── app/                    # Next.js App Router pages
│   ├── (frontend)/         # Public-facing pages
│   │   ├── courses/       # Course, chapter, lesson pages
│   │   ├── ask/           # AI chat interface
│   │   └── practice/      # Practice exercises
│   └── (payload)/         # Payload admin routes
│       └── admin/          # Admin panel
├── collections/            # Payload CMS collections
├── globals/                # Global configurations
├── hooks/                  # Custom hooks
├── ui/
│   ├── admin/             # Admin panel components
│   └── web/               # Frontend components
├── server/
│   └── payload/           # Payload server code
│       ├── collections/    # Collection configurations
│       ├── endpoints/     # Custom API endpoints
│       ├── jobs/          # Background job tasks
│       └── migrations/    # Database migrations
└── payload.config.ts      # Main Payload configuration

Key Features

Course Hierarchy

The learning content is organized in a hierarchical structure:

• Course: Top-level learning path
• Chapter: Thematic unit within a course
• Lesson: Individual teaching unit
• Exercise: Atomic learning item (questions, problems, quizzes)

AI Chat with Memory

The platform includes an intelligent chat system that:

• Maintains conversation context across sessions
• Stores long-term memories with vector embeddings
• Provides personalized tutoring based on learning history
• Supports PDF and image uploads for analysis

PDF to Exercise Pipeline

Convert PDF documents into structured exercises:

1. Upload PDF document
2. AI extracts content using Vision models
3. Content is chunked and stored in memory
4. Exercises are generated with context awareness

Multi-Tenant Architecture

Built-in support for multiple organizations:

• Tenant-scoped data isolation
• Configuration entries per tenant
• Locale support for courses

Collections

The platform includes the following Payload CMS collections:

Collection	Purpose
users	User accounts with roles
courses	Learning courses
chapters	Course chapters
lessons	Individual lessons
exercises	Practice exercises
conversations	Chat history
memoryItems	Long-term AI memory
media	File uploads
pages	Static pages
posts	Blog posts
prompts	AI prompt templates
tenants	Multi-tenant organization

Environment Variables

# Required
DATABASE_URL=mongodb+srv://...
PAYLOAD_SECRET=your-secret-key

# AI Services
GEMINI_API_KEY=your-gemini-key
OPENAI_API_KEY=your-openai-key

# Multi-tenant (when MCP_ENABLED=true)
DEFAULT_TENANT_SLUG=default
MCP_ENABLED=false

# Optional
MONGODB_MAX_POOL_SIZE=3
CRON_SECRET=your-cron-secret

# Best-effort telemetry target (sibling Admin app)
# Local Admin: http://localhost:3001
# Dev:  https://a-guy-admin-git-dev-aguy.vercel.app
# Prod: https://a-guy-admin.vercel.app
NEXT_PUBLIC_ADMIN_URL=http://localhost:3001

Development Commands

# Development
pnpm dev              # Start dev server
pnpm dev:full         # Clean restart with type generation

# Database
pnpm db:start         # Start local MongoDB
pnpm db:stop          # Stop local MongoDB
pnpm db:reset         # Reset database

# Code Generation
pnpm generate:types   # Generate Payload types
pnpm generate:importmap # Generate admin import map

# Testing
pnpm test:unit        # Unit tests
pnpm test:int         # Integration tests
pnpm test:e2e         # E2E tests with Playwright
pnpm test             # All tests

# Quality
pnpm typecheck        # TypeScript check
pnpm lint             # Linting
pnpm format:check     # Format check

# Release
pnpm release          # Semantic release

Documentation

Additional documentation is available in the docs/ folder:

• Platform Introduction - Unique approach, advantages, and architecture
• Course Hierarchy - Content structure patterns
• Exercises - Exercise types and rendering
• Exercise Import - PDF to exercise flow
• AI Services - AI integration details
• Access Control - Security patterns
• Block Rendering - Content block system
• Chat Context - AI memory architecture

AI Agent Support

This project includes AI-optimized documentation for autonomous agents:

• Pattern Index: .ai-docs/indexes/pattern-index.json - Code patterns by category
• Documentation Search: Fast keyword-based search
• Smart Doc Loader: Context-aware documentation loading

Generate indexes:

pnpm ai:generate-all

License

MIT