We built the POS we actually wanted

Three years ago, I bought my first POS for a 30-seat bistro in Bangkok. It cost ฿49,000 upfront, plus ฿2,400 monthly, plus 2.9% per card transaction. That was before the printer, the cash drawer, the iPad stand, the SIM card, the support contract. By month three, my actual cost-per-order was higher than my coffee margins.

Then the WiFi went down on a Saturday night. The POS didn't queue orders. It didn't degrade gracefully. It just stopped. We turned away ฿14,000 of dinner traffic in 90 minutes.

I remember sitting in the empty restaurant at 11 PM, scrolling through G2 and Capterra, looking for any system that would let me keep selling when the internet hiccupped. Every product page had "99.99% uptime" in the hero. None of them mentioned what happens at the 0.01%.

The compromise economy

I tried five other systems over the next year. Each asked me to compromise on something.

• The cheap one had a 1.2-second loading screen between every screen tap.
• The fast one charged 3.5% on every transaction · forever.
• The reliable one didn't support Thai language properly · my staff couldn't search menus.
• The pretty one was made for Western brunch cafés · couldn't handle Thai dish modifiers.
• The flexible one needed a software engineer to set up a single promotion.

Each one told me, in different words: "What you want is unreasonable. Pick two."

But I'm a restaurant operator. I don't pick two. My job is to serve everyone, every shift, every time. Picking two is what software people say when they don't know how to ship the third thing.

Five features. One philosophy.

So I started building. Not a product · just a tool, for my own kitchen. Three weeks of evening code. The first version was ugly. But it worked offline.

Then friends asked to use it. Then their friends. By month six, ten restaurants were running it. By month eighteen, a hundred. Today: 2,431.

The thing that surprised me most: the features I thought were table stakes turned out to be radical. Most POS systems still don't queue orders offline. Most still charge per-transaction. Most still hide their pricing behind "Contact sales."

Why ฿0 to start

Restaurant margins are thin. The first store is the riskiest. Most cafés in Thailand close within 24 months. Asking a new operator to pay ฿2,000/month before they've made their first ฿2,000 in profit is the same as asking them not to start.

We don't charge anything for the first store, under 100 orders/day, forever. Not a 14-day trial. Not a freemium tease. Free. As in: my first POS bill broke me, so this one won't break you.

When you grow past one store · or 100 orders/day · you'll have proven the business. Then we charge. Fairly. Transparently. With no transaction fees, ever.

What's next

This is Issue 06. Every quarter we publish one of these · engineering deep-dives, restaurant stories, things we got wrong, things we want to ship next. The next issue is in August.

If you're an operator reading this · thanks. We built this for you. If something's broken, email me directly: support.lemonpos@gmail.com.

If you're an investor reading this · we're not raising. The numbers work without it.

If you're another POS company reading this · please ship offline-first. We don't want to be the only ones.