Turn any topic into a structured, interactive course in seconds using Generative AI.
CourseForge is an intelligent LMS (Learning Management System) that uses a multi-tier AI agent architecture to generate comprehensive curriculums, lesson content, quizzes, and multimedia resources on the fly.
- Instant Outlines: Type "Advanced Python Patterns" and get a structured syllabus with modules and lessons in <10 seconds.
- Pro Mode (Reasoning): Uses DeepSeek/Llama-70B to plan complex topics with "Chain-of-Thought" reasoning for higher accuracy.
- Standard Mode (Speed): Uses Llama-8B for rapid, lightweight content generation.
- Multimedia Enrichment: Automatically fetches relevant YouTube videos and Unsplash images for every lesson.
- Audio Summaries: Generates "Podcast-style" audio summaries (Hinglish/English) using Neural TTS.
- Interactive Quizzes: Auto-generates MCQs to test knowledge retention at the end of every module.
- Live Code Execution: Includes an embedded Python sandbox (via E2B) that runs code in a secure micro-VM.
- Self-Healing Code: If the AI generates broken code, a background agent detects the error and fixes it automatically.
- Subscription Management: Full Stripe integration for credit packs and "Pro" monthly subscriptions.
- PDF Exports: Download entire courses or specific modules as professional PDFs.
- Sign Up: Login via Google/Auth0.
- Create: Click "New Course" and enter a topic (e.g., 'Quantum Computing for Babies').
- Customize: Choose Standard (Fast) or Pro (Deep Research) mode.
- Learn:
- Read the generated content.
- Run code snippets in the browser.
- Listen to the audio summary while commuting.
- Export: Click the "Download PDF" button to save your material offline.
Are you a developer, recruiter, or system architect? I have documented the entire engineering journey, from the high-level distributed system design to the specific prompts used for the AI agents.
Or jump to a specific section:
| Section | Description |
|---|---|
| 1. High-Level Architecture | System context, Microservices, and the Client-API-Worker triangle. |
| 2. The AI "Brain" | Deep dive into the ModelGateway, Task Tiers, and Self-Healing JSON loops. |
| 3. Event-Driven Workflow | How we use BullMQ & Socket.io for non-blocking, real-time generation. |
| 4. Database Design | MongoDB Schema, Hybrid Caching strategies (Redis + Mongo), and Polyglot Persistence. |
| 5. DevOps & Deployment | Docker strategies, Render pipeline, and Monorepo service orchestration. |
- Frontend: React Router v7, Chakra UI
- Backend: Node.js, Express, TypeScript
- AI Orchestration: OpenAI SDK (Custom Gateway), LangSmith (Tracing)
- Queue/Async: Redis, BullMQ, Socket.io (Redis Adapter)
- Database: MongoDB Atlas (Vector Search + Core Data)
- Infrastructure: Docker, Render, Stripe Connect
Built with ❤️ by Sumesh