From idea to production-ready product
🔗 Live platform: https://lessonvibe.com
Overview
LessonVibe is an AI-powered platform that enables teachers to create, customize, publish, and remix interactive ESL classroom games in seconds.
The project was built to solve a real classroom problem:
teachers want high-quality, level-appropriate speaking games, but they don’t have time to design them from scratch.
LessonVibe combines pedagogical design, AI prompt engineering, modern frontend architecture, and full DevOps automation into a single, scalable product.
What LessonVibe Does
LessonVibe allows teachers to:
- Generate ESL games using AI, calibrated to CEFR levels (A1–C2)
- Customize game mechanics, topics, and learner profiles
- Publish games with permanent shareable URLs
- Browse, remix, and reuse games created by the community
- Use the platform on desktop, tablet, or mobile
Supported game formats include:
- 🎲 Speaking Board Games
- 🧩 Jeopardy
- 🎭 Taboo
- 💰 Who Wants to Be a Millionaire
Why I Built It
This project reflects how I work at the intersection of:
- Learning design
- AI systems
- Product thinking
- Engineering quality
Rather than treating AI as a “content shortcut,” LessonVibe treats AI as a co-designer — with strong guardrails, fallback systems, and measurable outcomes.
It’s also intentionally built as a real product, not a prototype: authentication, usage limits, analytics, payments, testing, CI/CD, and documentation are all first-class citizens.
Technical Highlights
Frontend
- React 19 + TypeScript
- Vite for fast builds
- Responsive, mobile-first design
- Progressive Web App features
Backend
- Node.js + Express (TypeScript)
- RESTful API architecture
- Zod for runtime validation
- File-based persistent storage for published games
AI
- Anthropic Claude-3.5-Haiku
- Structured prompt engineering with CEFR calibration
- Full prompt and output tracking for quality analysis
Auth & Data
- Supabase (Google OAuth + Email)
- User tiers (Visitor / Free / Pro)
- Usage tracking and limits
- Admin dashboard with prompt analysis and exports
DevOps & Quality
- Dockerized deployment
- Fly.io production hosting
- GitHub Actions CI/CD
- Cross-browser Playwright testing
- Zero-downtime deployments with rollback
Product & Metrics Thinking
LessonVibe is instrumented from day one:
- Activation: second game created within 7 days
- Retention: 7-day return rate
- Completion: published games reaching classrooms
- Engagement quality tagging (poor → great)
- Pro conversion and MRR tracking
This allows the platform to evolve based on real usage, not assumptions.
Current Status
Upcoming roadmap items include:
- Advanced learning analytics
- LMS integration
- Offline exports
- Multi-language UI
- Mobile apps
Why This Matters
LessonVibe represents how I approach complex systems:
- Start with a real user problem
- Apply AI responsibly and measurably
- Design for scale, quality, and maintainability
- Ship production software — not demos
It’s a practical example of how AI, when paired with strong pedagogy and engineering discipline, can meaningfully improve learning experiences.
