⚠️ This project is actively under development. Things are evolving fast, some features are still being polished, and your feedback genuinely shapes the roadmap. Jump into Discussions to share ideas, report bugs, or just say hi.
Launchpad is a personalized coding education platform built on one core belief: learning to code shouldn't require an account, a subscription, or your data.
Tell it your career goal, and Launchpad generates a custom learning roadmap pulling from a massive on-device curriculum — then walks with you through lessons, quizzes, projects, mock interviews, and even helps you build a resume at the end. All of it runs 100% in your browser. No servers tracking you. No sign-ups. No catch.
| 38 Tracks | 797 Lessons | 7,221 Quiz Questions | 207 Projects |
| 1,980 Daily Challenges | 43 Achievements | 232 Interview Questions | 0 Accounts Required |
| Dashboard Personalized overview with stats, career readiness, and one-click lesson re-entry |
Lesson View (Reading-First) Documentation-style layout with hidden outline and focused reading column |
 |
 |
| Course Outline (Toggle) Slides in as an overlay when you need navigation; stays hidden while reading |
Mobile Dashboard Touch-friendly bottom nav, adaptive cards, safe-area aware FAB |
 |
 |
- ✨ Core Features
- 🧠 How Personalization Works
- 🛠️ Tech Stack
- 🚀 Getting Started
- 📖 Course Catalog
- 📜 Certificates
- 🔒 Privacy, By Design
- 🏗️ Architecture Overview
- 🗺️ Roadmap
- 🤝 Contributing
- 📜 Changelog
🧠 Personalized Roadmaps
Your personalized learning path is generated instantly by Launchpad's built-in deterministic engine — right in your browser. No API keys, no waiting, no network calls. It combines your career, languages, skill level, and availability into a structured plan.
- 12-point validation on every generated roadmap (phase counts, lesson references, sequencing, and more)
- Variable-length roadmaps (4–10+ phases) that scale with your goals and language count
- Auto-injected prerequisite languages with clear 'Required for' labeling
- A dedicated AI-focused bonus phase near the end of every track
- Every roadmap task links straight into the matching lesson
📚 Learn Tab — Lessons Across 38 Technologies
From Python and JavaScript to Rust, Swift, React, and PostgreSQL — each track has in-depth lessons. Every single lesson includes:
- A "why it matters" framing + prerequisites
- Multiple worked code examples
- Common pitfalls & real-world applications
- Collapsible interview questions
- A mini project + 10-question quiz with explanations
- A curated, privacy-respecting YouTube tutorial (youtube-nocookie.com)
📖 Documentation-Style Reading Experience
Lessons read like professional documentation — clean typography, clear section hierarchy, comfortable reading width, and a collapsible course outline sidebar that shows your progress through the track. No box-heavy layouts, no visual clutter. Just content.
✏️ Inline Code Editor — Edit & Run Everywhere
| Language type | How it runs |
|---|---|
| JavaScript / TypeScript | Sandboxed iframe, no eval, 5s timeout, network APIs stripped |
| HTML / CSS | Instant live preview |
| Python | Pyodide (Python via WebAssembly) |
| SQL | sql.js in-browser, with DB Fiddle for Postgres-specific features |
| Bash / Shell | Simulated shell with a fake virtual filesystem |
| Compiled languages (Java, C++, Go, Rust, etc.) | One-click launch into Replit / OneCompiler / official playgrounds |
| Frameworks (Svelte, Vue, Angular, Node) | Direct links to official playgrounds / StackBlitz |
🤖 AI Tutor, Mock Interviews & Code Review (Bring Your Own Key)
No platform-funded AI costs here — you plug in your own free or paid API key (Gemini, Groq, OpenRouter, OpenAI, Anthropic, or any custom OpenAI-compatible endpoint), and unlock:
- 💬 AI Tutor — full conversational help, multi-chat history, all stored on-device
- 🎯 Mock Interview Mode — a simulated senior technical interviewer asks questions one at a time, scores you, and tells you exactly what to study next
- 🔍 AI Code Review — paste shipped project code and get a structured review with a score out of 10
📜 Certificates & Career Readiness
- Per-language certificates unlock once you finish every lesson in a track and hit a 75%+ quiz average
- Career Master Certificate unlocks at a 100% Career Readiness Score — a weighted blend of roadmap progress (30%), quiz performance (30%), projects shipped (20%), and interview practice (20%)
- Every certificate has a public, privacy-respecting verification page at
/verify/LP-XXXXXXXX
🔁 Spaced Repetition & Study Tools
- SM-2 spaced repetition (the same algorithm Anki uses) tracks every quiz question you get wrong and surfaces it for review at the scientifically optimal interval
- Flashcards tab — auto-generated from lesson key concepts, interview questions, and quiz answers (14,385 potential cards)
- Weak Areas card on the Learn tab shows your top missed questions with one-click "Review Now" deep-links
- Quiz review mode — choose between "Take fresh quiz" or "Review difficult questions"
- Lesson bookmarks — star lessons to revisit later
- Read time estimates on every lesson
- Print-friendly lesson view — one-click print to PDF
- "I don't understand" button on every quiz question — sends context to the AI Tutor
🎮 Gamification, Community & Daily Habits
- 43 achievement badges across 4 rarity tiers (Common, Rare, Epic, Legendary)
- 10-level XP curve that rewards lessons, quizzes, projects, streaks, interviews, flashcards, and bookmarks
- 1,980 daily challenges across all 38 languages, rotating weekly per user
- A built-in Community tab (GitHub Discussions via Giscus) — Announcements, Help & Questions, Show & Tell, General Chat, and Feature Requests
- Calendar with recurring study sessions, reminders, and snooze support
- Installable as a PWA, with offline support and a mobile-first bottom nav
flowchart TD
A[Onboarding: Your Goals & Background] --> B[Roadmap Generation (Built-in Engine)]
B --> C[Learn Tab: 797 Lessons]
B --> D[Daily Challenges Pool]
B --> E[Projects: 207 Available]
C --> F[Quizzes + Spaced Repetition]
E --> G[Ship Project → AI Code Review]
F --> H[Career Readiness Score]
G --> H
D --> H
H -->|100%| I[🏆 Career Master Certificate]
H -->|per track| J[📜 Language Certificates]
| Layer | Technology |
|---|---|
| Framework | Next.js 16 (App Router) + TypeScript 5 |
| Styling | Tailwind CSS 4 + shadcn/ui (Liquid Glass design system) |
| State | Zustand, persisted to localStorage |
| Roadmap Engine | Built-in deterministic engine (100% on-device, no API key needed) |
| Tutor / Interview / Review AI | BYOK — Gemini, OpenAI, Anthropic, Groq, OpenRouter, or custom endpoint |
| Code Execution | Sandboxed iframe, Pyodide, sql.js, simulated shell |
| Certificate Verification | Supabase (Postgres + RLS) |
| Community | Giscus (GitHub Discussions) |
| Syntax Highlighting | react-syntax-highlighter (Prism, vscDarkPlus) |
| Testing | Vitest (70 tests across 4 suites) |
v6.000 architecture: Every lesson and quiz question has a permanent stable identity ("slug") decoupled from its display order. Lesson metadata is build-generated from content, never hand-maintained.
v6.002 content pipeline: All 797 lessons are authored as Markdown files in
content/{track}/and compiled to per-track JSON at build time. The runtime loads only the track you open (~200-470 KB) instead of the entire curriculum.
v6.008 UX redesign: Documentation-style lesson reading, collapsible course outline sidebar, stronger glass contrast, and a dashboard "Continue your lesson" card for one-click re-entry.
1. Clone & install
git clone https://github.com/dumzvybez/launchpad.git
cd launchpad
bun install2. Set up environment variables — create .env.local:
# Optional — only used if you self-host and want platform keys for AI Tutor (BYOK is the default)
# Certificate verification (Supabase) — v5.77+
NEXT_PUBLIC_SUPABASE_URL=https://your-project.supabase.co
NEXT_PUBLIC_SUPABASE_ANON_KEY=your_anon_key
SUPABASE_URL=https://your-project.supabase.co
SUPABASE_SERVICE_ROLE_KEY=your_service_role_key # server-only, never expose
# Self-hosting — v5.77+
NEXT_PUBLIC_APP_URL=https://your-domain.com
# Certificate forgery protection — v5.77+ (optional but recommended)
CERT_SECRET=your_random_secret_stringFree keys: Gemini · Groq · OpenRouter · Supabase
Roadmap generation needs no keys — it runs entirely in your browser. AI Tutor, Mock Interviews, and Code Review are BYOK (bring your own key).
3. Run it
bun run devThen open http://localhost:3000 🎉
38 technologies, each with ~21 lessons today (~797 lessons total). The v6.000 architecture supports scaling to 100-150+ lessons per track — every lesson now has a permanent stable identity ("slug") so tracks can be reorganized and expanded without ever breaking user progress.
| Python | JavaScript | TypeScript | HTML |
| CSS | SQL | Java | C |
| C++ | C# | Go | Rust |
| Swift | Kotlin | PHP | Ruby |
| R | Dart | Bash | React |
| Next.js | Django | FastAPI | Flask |
| Svelte | Vue | Angular | Node.js |
| PostgreSQL | MongoDB | Docker | Tailwind |
| Express | GraphQL | Kubernetes | Terraform |
| PyTorch | TensorFlow |
Each track represents roughly 50–200 hours of learning time, beginner to advanced.
| Per-Language Certificate | Career Master Certificate | |
|---|---|---|
| Unlocks when | All lessons in a track complete + 75%+ quiz average | 100% Career Readiness Score |
| Format | Landscape PDF, deterministic ID (LP-XXXXXXXX) |
Gold-accented PDF, ID prefix LP-CAREER-XXXXXXXX |
| Verification | Public page at /verify/LP-XXXXXXXX |
Same verification system |
No accounts. No servers storing your data. No analytics. Everything lives in your browser's localStorage.
- Your AI API keys never leave your device
- Chat history with the AI Tutor stays local
- Roadmap generation runs entirely on your device — nothing is sent anywhere
- The Community tab requires a GitHub account to post, but your Launchpad progress is never shared there
launchpad/
├── content/ # Markdown lesson source (797 files across 38 tracks)
│ ├── python/ # 21 lessons
│ ├── javascript/ # 21 lessons
│ └── ...
├── public/content/ # Compiled per-track JSON (lazy-loaded at runtime)
├── scripts/
│ ├── gen-lesson-meta.ts # Markdown → metadata (slugs, counts, track maps)
│ └── compile-content.ts # Markdown → per-track JSON
├── src/
│ ├── app/ # Next.js App Router (/, /verify/[id], /api/*)
│ ├── components/
│ │ ├── glass/ # Liquid Glass design primitives
│ │ ├── learning/ # Lesson experience (sidebar, header, blocks)
│ │ ├── shell/ # AppShell, Sidebar, TopBar, Onboarding
│ │ ├── views/ # 18 main views (Dashboard, Learn, Projects, etc.)
│ │ └── ui/ # shadcn/ui components
│ ├── lib/
│ │ ├── curriculum/ # v6.004 module catalog, assessments, capstones
│ │ ├── store.ts # Zustand store (persisted to localStorage)
│ │ ├── personalization-engine.ts # On-device roadmap generation
│ │ ├── sm2.ts # Spaced repetition algorithm
│ │ └── certificate-utils.ts # HMAC-signed certificate IDs
│ └── data/ # YouTube links, career data
├── supabase/schema.sql # Certificate verification table + RLS
└── tests/ # Vitest suites (content, curriculum, certificates, store)
Content pipeline: content/*.md → scripts/compile-content.ts → public/content/*.json → lazy-fetched by content-loader.ts when a user opens a track.
Identity system: Every lesson has a permanent slug (e.g., python-variables-and-data-types) decoupled from its positional ID. All persisted user state (progress, bookmarks, flashcards, certificates) uses slugs — so lessons can be reordered or expanded without breaking anything.
- ✅ v6.000–v6.006: Stable identity system, per-track lazy loading, dead code removal, TypeScript hardening, test suite
- ✅ v6.010: SEO & social preview fixes, reading-first lesson redesign, liquid glass readability pass, mobile UX overhaul, verified-statistics README
- 🔄 Next: Populate forward-looking curriculum fields (aiContext, learningObjectives, skillsTaught, practiceChallenges) on lessons
- 🔄 Next: Wire the v6.004 graduated assessment ladder (module quizzes, checkpoint exams, certificate exams)
- 📋 Future: RAG/vector search for AI tutor, leaderboards
Pull requests, issues, and ideas are all welcome.
- Repo: github.com/dumzvybez/launchpad
- Discussions: Share feedback, report bugs, request features
- Live App: launchpadedu.vercel.app
- Developer: Dumindu Dulara Wanasinghe — Portfolio
- SEO & social previews: Fixed OG image dimensions (1734×907 → 1200×630), optimized file size (1.5 MB → 265 KB), corrected lesson counts in metadata (630 → 797), enriched JSON-LD structured data
- Reading-first lesson redesign: Course outline hidden by default; opens as a smooth slide-in overlay via a clear “Open course outline” toggle. Removed duplicate lesson title. Added dedicated prev/next lesson navigation. Wider, centered reading column (max-w-3xl/4xl) — Apple/Stripe docs quality
- Liquid glass readability: Increased glass tint and popover opacity (light & dark) so dropdowns, tooltips, and modals are near-opaque for WCAG-AA text contrast. Darkened muted-foreground for secondary text
- Mobile UX overhaul: Fixed high-contrast bottom nav with safe-area support, repositioned AI Tutor FAB to clear the nav bar, ≥44px touch targets, bottom-sheet patterns for notifications/version dialog/command palette, hidden desktop-only keyboard hints on touch devices
- Dashboard & shell polish: Cleaned welcome-header hierarchy, neutralized the 0% career-readiness color, aligned stats row, fixed text overflow with line-clamp, shortened top-bar search placeholder, reduced redundant sidebar active-state highlighting
- Fixed React warning: Moved
history.replaceStateout of render body intouseEffect(Rules of Hooks compliance) - Verified statistics: README counts read directly from the codebase (797 lessons, 7,221 quiz questions, 207 projects, 1,980 challenges, 43 achievements, 232 interview questions)
- Documentation-style lesson reading (no box-heavy layout, clean typography hierarchy)
- Collapsible lesson sidebar with proper compact rail + expand button (persists preference)
- Liquid glass readability fixes (onboarding dropdowns, notification center, stronger tint)
- Dashboard "Continue your lesson" card for one-click re-entry
- Screen size optimization (tighter dashboard spacing for first-viewport fit)
- Updated all project links to new URLs
- Sticky lesson sidebar with track progress
- Constrained reading width (max-w-3xl)
- Glass tint opacity increased for readability
- Wider main content (max-w-7xl) for 1440px+ screens
- Removed 11 MB dead content bundles
- Removed unused Prisma setup
- TypeScript ignoreBuildErrors removed — all errors fixed
- Added Vitest test suite (70 tests)
- Fixed stale documentation
- Stable lesson identity (slugs)
- Per-track lazy content loading
- Scalable curriculum architecture (modules, capstones, assessments)
- Guided lesson experience components
Built solo, built free, built for everyone learning to code. 💙