When an app stops being a 30-day build: Let's Raise It

The whole premise of 30Apps is speed — one build, one day, move on. But every so often a build refuses to stay small. Let's Raise It was one of those.

It started as a simple idea: school fundraising platforms charge 20%+ in fees, which is absurd. Could I stand up a transparent 2–5% alternative? The first version came together fast. Then the real work showed up: Stripe Connect payouts, multi-tenant orgs, TCPA compliance, 10DLC, donor flows, accounting emails on a cron.

What agents were great at

• Scaffolding the boring 80% — schema, RLS policies, email templates, CRUD.
• Holding context across a big codebase once I gave them the right anchors (a context model, personas, a job-to-be-done doc).

What still needed me

• The judgment calls. Which fee structure, which compliance corners are non-negotiable, what "done" means for a payout flow handling real money.
• Production hardening. Environment guards, cron secrets, staging vs prod separation — the unglamorous stuff that decides whether a thing can actually take customers.

The lesson: agents collapse the cost of the first 80%, which means you spend your attention on the 20% that was always the hard part. That's not less work. It's better-targeted work.