GrowGo

Problem

Local food is stuck in cash and folding tables. A farmers market runs a few hours a week — miss the window and both sides lose. Vendors juggle cash boxes and Venmo with no pre-orders, no inventory visibility, and no way to be found off-market. Buyers who want fresh, local food have no marketplace built for the people who actually grow it.

GrowGo is the digital storefront for home-grown goods: buyers discover vendors and markets on a map, pre-order and pay in-app, then pick up with a QR code. Vendors get a year-round storefront, messaging, and verified-purchase reviews.

What happened in 2025

I pitched it for VC funding as the app the farmers-market world was missing. The love was universal; the checks never came. The plan didn’t help — Rust and Go microservices, native iOS, native Android, and a web app, all specced before a single vendor was onboard. Ambitious enough to swing for the fences on other people’s money, and exactly backwards. I archived it here with an honest postmortem and moved on.

The resurrection

Moving on didn’t take. In August 2026 I brought it back with the order inverted: build the working thing first, then talk to investors.

The rebuilt stack is sized for a solo founder instead of a fantasy org chart. One Flutter codebase covers iOS and Android — every screen of the buyer flow is clickable, from the vendor map through cart, checkout, and QR pickup. Behind it sits a Cloudflare Workers + D1 backend carrying the full marketplace schema: vendors, products, market check-ins, one-vendor-at-a-time orders with server-side pricing and atomic stock decrements, messaging, and reviews gated behind a verified purchase. The pitch got the same treatment — rebuilt from scratch around the working prototype instead of mockups.

Constraints

Pre-revenue infrastructure has to cost approximately nothing, which is why the backend is Cloudflare-first with standard fetch and standard SQL as the escape hatch. Marketplace trust has to be structural, not moderated after the fact — reviews require a real order, and prices are computed server-side. And spring market season is the real deadline: launch after it starts and the year is gone.

Current state

Deliberately not launched. What remains is the wiring with real-world keys in it: Stripe payments and Connect onboarding, the production Cloudflare deploy, and image storage. Until those land, the pitch is the public face — it lives on this site behind a numeric keypad at /misc/growgo. If you’re the kind of person who should see it, ask me for the code.

What I’d do differently

It’s what I am doing differently: product before pitch. The 2025 version asked strangers to fund intentions. The 2026 version opens with an app you can tap through and a backend you can hit with curl, and asks a much smaller question.