Скачать или смотреть AI-Built E-Learning Platform (Modular Monolith): Laravel, Next.js, NestJS, Stripe, Docker, Redis

I miei corsi: https://www.hidran.it
my ebooks: https://www.hidran.it/hidran-libri-pr...
In this end-to-end build, I show how I’m designing and shipping a real e-learning platform — courses, subscriptions, payments, video lessons, attachments, teacher collaboration, and an admin area — using AI as my co-dev. You’ll see both paths: (1) generating the first working app with Lovable + Stripe test mode and (2) developing locally with a monorepo (Next.js web + NestJS API), Dockerized infra, seeds, tests, and a clean modular architecture.

What you’ll learn

• Prompt-to-product workflow with ChatGPT/Gemini (custom instructions, role setup, project plan/progress logs).
• Modular monolith with extractable services and clear domain modules.
• Identity & Access: Auth, roles/permissions (RBAC/ACL).
• Catalog: categories → courses → sections → lessons; tags; search & filters.
• Learning: enrollments, progress, notes, quizzes, lesson assets (PDF, ZIP, PPT, video).
• Commerce: pricing, coupons, subscriptions, invoices, revenue share, teacher collaboration.
• Media pipeline: uploads, transcoding (HLS), storage (S3/MinIO), CDN, watermarking.
• Infra & DevX: Docker Compose, Postgres, Redis, MailHog, seeds/fixtures, .http endpoint tests.
• Observability & Ops: OpenTelemetry, CloudWatch; GDPR & EU VAT awareness with Stripe.
• Frontend UX: search, category filters, course viewer, login/sign-up, dashboard for students/teachers/admin.

API outline (sample)

Auth → /auth/register, /auth/login, /auth/refresh
Catalog → /categories, /courses, /courses/:id, /sections, /lessons
Learning → /enrollments, /progress, /notes, /quizzes
Commerce → /prices, /checkout, /subscriptions, /invoices, /coupons
Collab/Admin → /teachers, /collaborators, /admin/*

Tech stack shown

• Lovable for first working prototype (free daily credits) connected to Stripe sandbox; purchase flow verified.
• Local monorepo: Next.js App Router (web) + NestJS API; per-feature tests (integration/spec), endpoint smoke tests via *.http files.
• Laravel option (my usual stack) for an alternative backend; PHP 8.3/8.4; Cashier (Stripe), Spatie Permissions, Horizon, queues.
• Storage/infra: MinIO (S3 compatible), Redis cache/queues, Postgres, Docker Compose; AWS targets like Fargate/Elastic Beanstalk.
• Media: HLS transcoding pipeline with CDN + optional watermarking; asset types: pdf/zip/video/ppt.

Data model (high-level)

Users, Roles, Permissions → Teachers/Students (RBAC)
Categories ↔ Courses → Sections → Lessons → LessonAssets (pdf/zip/video/ppt)
Enrollments, Subscriptions, Prices, Coupons, Invoices, RevenueShare
Collaboration (co-authors), Notes, Quizzes, Progress

Productivity system I use

To keep momentum across LLM sessions I maintain three files in the repo:

project.md — the master prompt/spec the AI reads each time.

project-plan.md — phases & features the AI should complete.

project-progress/step-log.md — a commit-style log the AI updates on every change.
If I switch models or run out of credits, the next session just “reads the state” and continues — same context, less friction.

Gotchas I hit (and fixes)

• Docker boot order: ensure Postgres is ready before migrations/seeds (health checks + wait-for scripts).
• Seeded data: rebuild volumes to re-run bootstrap and get fresh fixtures.
• Teacher console crash: reproduced locally, then handed to an LLM which patched routing/state and restored the dashboard.
• Stripe test keys: store securely; use webhooks/test clocks for subscription flows.
• MinIO/S3: set correct CORS for signed uploads and HLS segment serving.

Who is this for?

Indie devs, SaaS builders, and teams shipping course marketplaces or training portals who want a modern, scalable foundation without reinventing auth, payments, and media. The video’s narration/instructions are in English; my IDE/UI may show Italian defaults at times.

Try/replicate (mini checklist)

[ ] Create your role-prompt (software architect) + custom instructions
[ ] Generate the initial app in Lovable; connect Stripe test keys
[ ] Run Docker Compose (Postgres, Redis, MailHog, MinIO)
[ ] Seed data and verify migrations
[ ] Hit endpoints with the provided *.http files
[ ] Add tests per feature; track every commit in your step log
[ ] Compare Lovable vs local dev for speed/control

If you find this useful, drop a like, comment what you want built next (quizzes? HLS watermarking? multi-tenant? marketplace payouts?), and subscribe for the follow-up where I push the same spec with ChatGPT generating a Laravel backend end-to-end.