Skip to content
nyuchi-docs

Deployment overview

Nyuchi runs production workloads on three substrates: Cloudflare (edge, Workers, R2, D1, KV, Hyperdrive), Vercel (Next.js apps and the public marketing surface), and Supabase (Postgres, Auth-adjacent storage, edge functions where Cloudflare isn’t the right fit). This section captures the patterns we’ve already paid for so a new project doesn’t reinvent them.

What needs to land here

Section titled “What needs to land here”

Cloudflare

Section titled “Cloudflare”
  • The Workers project layout we use and the wrangler.toml conventions.
  • D1 migrations: where they live, how we apply them, branch safety.
  • R2 bucket naming and lifecycle rules.
  • KV namespace partitioning per env.
  • Hyperdrive in front of Supabase Postgres — when, why, the failure modes.
  • Edge / origin routing rules at the zone level.

Vercel

Section titled “Vercel”
  • Project naming, env-var doctrine, preview-deployment hygiene.
  • The Next.js version we standardise on and the App Router conventions.
  • Image / asset handling, edge runtime choices.
  • How we wire docs.nyuchi.com and platform.nyuchi.com once DNS is set up.

Supabase

Section titled “Supabase”
  • Project-per-environment layout, the migration workflow, branch databases.
  • RLS doctrine: what we allow, what we forbid, how we audit it.
  • Edge functions: when to reach for them vs. a Cloudflare Worker.
  • Realtime: opt-in only, with explicit channels.

Cross-links (once content lands)

Section titled “Cross-links (once content lands)”
  • nyuchi/mukoko-platform — reference deployment for the Console.
  • The internal infra/ repo — Terraform / wrangler configs (not public).