Shipped Akamaru — a SaaS boilerplate I extracted from two apps already running in production

After shipping two SaaS apps on the same stack, I got tired of rebuilding

auth, billing, admin, and audit trails every time. So I extracted the

base into a boilerplate and hardened it until my own production apps run

on it today.

One of those apps — Opsrift — is live at opsrift.com. Every feature in

Akamaru earned its place by solving a real problem there first.

What's wired on day one:

• Clerk auth (OAuth, MFA, org seats)

• Paddle billing (MoR — no sales tax headaches)

• Admin panel with user search, refunds, impersonation

• Postgres + Drizzle, tenant-scoped queries

• Claude Sonnet 4.6 streaming + daily spend cap

• GDPR export + audit logs

• 101 tests + CI + Vercel deploy

Three design decisions I keep having to defend:

1. Not Next.js. React + Vite + Express 5 — fewer moving parts, no

framework lock-in, any host works.

2. Paddle not Stripe. Merchant-of-Record means sales tax is handled

globally. If you're outside the US, this alone is worth the price

of admission.

3. Claude Sonnet 4.6 wired with a daily spend cap so a bug can't drain

your Anthropic bill.

$149 launch / $199 after / MIT license / 30-day refund.

Would love feedback from anyone who's sold a boilerplate before — especially

on pricing, positioning, and whether the Paddle angle lands with non-US

buyers.