replica-architect

1# replica-architect
2 
3Reads `replica/recon.md` and `replica/features.csv`. Writes
4`replica/architecture.md` (template: architecture.md in this folder).
5 
6If there is no recon map, stop and run `/replica-recon` first. Planning a
7clone from memory of what an app does is how you miss half of it.
8 
9## Step 1: the stack
10 
11Use what the user already knows if they have a stack. Otherwise the default,
12because every part is managed, documented and cheap at zero users:
13 
14| layer | default | swap for |
15| --- | --- | --- |
16| web app | Next.js (App Router) + TypeScript | Remix, SvelteKit, Rails |
17| styling | Tailwind, tokens from replica-design | CSS modules |
18| mobile | Expo (React Native) | SwiftUI, Kotlin |
19| database | Postgres on Supabase or Neon | PlanetScale, SQLite (Turso) |
20| ORM | Drizzle or Prisma | raw SQL |
21| auth | Supabase Auth or Auth.js | Clerk |
22| payments | Stripe Checkout + Billing | Lemon Squeezy, Paddle |
23| email | Resend or Postmark | SES |
24| jobs | Vercel Cron, Inngest or Trigger.dev | a worker on Fly |
25| files | Supabase Storage or Cloudflare R2 | S3 |
26| hosting | Vercel | Netlify, Fly, Render |
27 
28Write each choice with one line of why. One database. No microservices. The
29clone does not need the original's architecture, it needs the original's
30features.
31 
32## Step 2: the schema
33 
34Turn the inferred data model into SQL. For every table:
35 
36- `id uuid primary key default gen_random_uuid()`, `created_at`, `updated_at`
37…