La réalité full-stack en production
What vibe coders think full-stack is, a frontend on top of a backend, two blocks, versus what production actually demands: thirteen stacked layers. Frontend, APIs and backend logic, database and storage, auth and permissions, hosting and deployment, cloud and compute, CI/CD and version control, security and RLS, rate limiting, caching and CDN, load balancing and scaling, error tracking and logs, availability and recovery. The concept names the gap between "it runs on my prototype" and "it holds when 5,000 people show up at once".
Strengths
- Gives a mental checklist before you say "it's production-ready", you know what you haven't covered
- Reveals what your host (Vercel, Supabase, Render) already absorbs, so you don't reinvent or pay twice
- Shared vocabulary with a senior dev or ops, you can name precisely what is missing
Limitations
- Opposite risk: believing all thirteen layers are required at MVP and over-engineering a three-user project
- The list is a map, not a recipe, it doesn't tell you WHEN each layer actually becomes necessary
Best for
- Vibe coders who shipped a prototype and want to understand what separates them from real production
- PMs/POs scoping an AI project who must estimate the real workload beyond the screen
- Designers and no-code makers moving to build who discover what is behind the curtain