Why This Stack
The goal was to ship, not to architect. Every tool in this stack exists because it collapses something that used to require either a team or a week of setup. Supabase replaces a Postgres instance plus auth plus storage plus realtime. Vercel removes the deployment mental model entirely. Resend means transactional email doesn't require configuring an MTA. The stack is optimized for a one-person operation where the bottleneck is always feature work, not infrastructure.
The Stack
| Tool | Role | Cost / Tier |
|---|---|---|
| Next.js 15 | Full-stack React framework — app router, API routes, server components | Free |
| Supabase | Postgres database, auth, storage, realtime | Free tier → $25/mo Pro |
| Vercel | Deployment, edge functions, analytics | Hobby free → $20/mo Pro |
| Resend | Transactional email API | Free tier → $20/mo |
| Stripe | Payments, subscriptions, invoicing | 2.9% + 30¢ per transaction |
| Linear | Issue tracking, roadmap | Free for solo → $8/mo |
Notes on Each Choice
Next.js: The app router took adjustment but the server component model genuinely reduces client-side code. Worth the learning curve if you're starting fresh in 2026.
Supabase: The dashboard is good enough that I've written zero custom admin tooling. Row-level security took a day to understand. It was worth understanding.
Stripe: The Stripe Billing API is more complex than the pricing docs suggest. Budget a week for subscription flows, not a day.
I'd evaluate PlanetScale or Neon as Supabase alternatives earlier. Supabase's connection pooling at scale has surprised me with limitations I didn't anticipate at signup. Also: I'd set up Linear earlier. I delayed it because it felt like overhead for a solo project. It wasn't.