The Boring Stack That Ships (And Sells)

Stop researching tools. Pick this stack and build.

You're not building Snowflake. You're building a small app that needs to make $2K MRR so you can list it for $40K to $70K on Birexit. The stack you pick should optimize for one thing: a buyer can take it over on day one without crying.

Buyers don't pay a premium for clever. They pay a premium for boring. Boring is what they can read, fork, and run with.

The Cheat Sheet

Total at zero traffic: $0. Total when you have 100 paying users: $60 to $85/month. That's it.

Why Cursor Beats Bolt for the Long Run

Cursor is just VS Code with a great AI sidebar. It feels boring. That's the point.

The file structure is real. Your code lives on your disk. When a buyer asks "can I download the repo and run it locally?" the answer is yes, immediately, no platform export dance. Bolt and Lovable are fine for the Friday-night MVP, but the apps that close at 3.5x to 4x multiples on the marketplace tend to be Cursor-built Next.js repos that any halfway-decent dev can pick up.

Real Talk: I built the first version of one app in Bolt over a weekend, validated it, then rebuilt in Cursor over the next two weekends. The Cursor version is the one that sold.

Supabase: The Backend You Don't Have to Explain

Supabase gives you:

• ·A Postgres database (the most boring, most portable DB on earth)
• ·Auth with email, magic links, Google, GitHub (you don't write a single auth bug)
• ·Row-level security so you don't accidentally leak User A's data to User B
• ·A free tier that gets you to your first 50,000 monthly active users

Why not Firebase? Because Firebase is Google's proprietary NoSQL flavor, and buyers wince at vendor lock-in. Postgres is the universal language. If a buyer wants to move off Supabase later, they pg_dump and walk away. Try that with Firebase.

Why not your own Postgres on Railway / Fly / Render? Because you'll spend three weekends building the auth flow Supabase ships out of the box. Don't.

Stripe: There Is No Second Choice

Stripe Checkout takes about 30 minutes to wire up with Cursor's help. It handles cards, Apple Pay, Google Pay, taxes (with Stripe Tax bolt-on), subscriptions, refunds, customer portal, dunning emails when cards fail. All of it.

Lemonsqueezy and Paddle are fine for solo SaaS, but every buyer on Birexit has integrated Stripe before. Zero of them want to learn your custom payment vendor. Just use Stripe.

Tip: turn on Stripe Tax from day one. The $0.05 per transaction is nothing compared to the headache of backfilling VAT collection when your app gets traction in the EU.

Vercel: Deploy in 90 Seconds, Stop Thinking About It

Push to GitHub, Vercel deploys. That's the loop. Preview URLs for every PR, automatic HTTPS, edge network, no DevOps.

You'll outgrow the free tier when you cross ~100GB bandwidth or start getting serious traffic. That's a $20/month problem you should be thrilled to have.

The buyer angle: every technical buyer has used Vercel. They'll see the deploy in the GitHub Actions tab, recognize the pattern in 30 seconds, and not have a single question about how the site goes live.

Resend: Transactional Email That Doesn't Suck

Resend is built by the React Email people. The API is two lines. The dashboard is clean. The free tier (3,000 emails / month) covers you through your first 500 users easily.

Why not SendGrid or Mailgun? Because you'll spend an hour fighting the DNS docs, then another hour figuring out why your emails land in spam. Resend's onboarding hand-holds you through DKIM/SPF/DMARC in about 10 minutes.

Use it for: welcome emails, password resets, receipt emails, "your trial ends in 3 days" nudges. Not for marketing blasts (use Loops or ConvertKit for that, separate thing).

What This Stack Costs at Real Traffic

Let's say you've got 200 paying users at $19/month. That's $3,800 MRR, healthy enough to list.

• ·Cursor: $20/mo (you, one seat)
• ·Supabase Pro: $25/mo
• ·Stripe: ~$120/mo in fees on $3,800 revenue (~3%)
• ·Vercel Pro: $20/mo
• ·Resend: $20/mo (covers ~50K emails)

Total infra: about $85/month plus Stripe fees. Gross margin on a $19/mo SaaS at this scale is comfortably above 85%. Buyers love that math.

What to Skip (For Now)

Don't bolt on these until you actually need them:

• ·Posthog / Mixpanel. Vercel Analytics is free and good enough for the first 100 users.
• ·Sentry. Supabase logs + Vercel logs catch 90% of bugs. Add Sentry at month 6 if you must.
• ·A CDN like Cloudflare. Vercel's edge is already a CDN. Don't double up.
• ·A separate auth provider (Clerk, Auth0). Supabase Auth is fine. Clerk is great but it's another $25/mo, another vendor in the diligence doc, and another login flow for the buyer to learn.

The buyer will read your package.json. The shorter that file, the higher your multiple.

The Real Reason This Stack Wins

It wins because every single tool here is something a buyer has used before. There are no surprises in the diligence call. There's no "wait, what's this Convex thing?" There's no "you self-host the database? in a Docker container? where?"

The boring stack is a feature. Ship on it. List on it. Sell.

Pick the stack. Open Cursor. Build.