He shipped in days and made $22k in a week

In early March 2023, OpenAI released the ChatGPT API. Most developers said "cool, I'll add it to the backlog." Tony Dinh — an indie hacker from Vietnam — opened his laptop.

Within about 5 days of the announcement, he'd shipped TypingMind: a fast, feature-rich interface for the ChatGPT API (search, folders, custom prompts, multi-model support). The first working version took him roughly a day to build. It made around $10k in the first few days, and about $22k in the first week after a #1 Product Hunt launch. It's since crossed $1M+ in lifetime revenue and reportedly runs at ~$130k+/month.

No team. No funding. Let's break down what actually made it work — including one pricing decision most people miss.

Move 1: Ship the obvious better thing in days

The raw ChatGPT interface was clunky. Tony's insight wasn't invention — it was noticing that a brand-new platform had just been born and needed better tools immediately. He wasn't first to think "someone should build a nicer UI for this." He was first to ship it.

When a big new platform drops — an API, a model, a marketplace — there's a short window where "the obvious better version" is wide open. The prize goes to whoever ships in days, not months.

The speed wasn't luck. Tony had spent ~2 years as a solo indie hacker before this, building tools like DevUtils, Xnapper, and BlackMagic (plus failed ones). He could move fast because he'd already done the reps.

Move 2: The pricing trick nobody screenshots — BYOK

Here's the part that's genuinely clever and rarely talked about. TypingMind used a "bring your own key" (BYOK) model: you pay Tony a one-time license fee for the interface, then plug in your own OpenAI API key and pay OpenAI directly for usage.

Think about what that does:

• Near-zero running costs for Tony. He's not paying the AI bills — you are. Most AI startups get crushed by API/GPU costs; TypingMind sidestepped that entirely.
• No subscription friction. A one-time payment is an easier "yes" than yet another monthly subscription, especially for developers.
• Aligned incentives. Heavy users don't cost him more; they just pay OpenAI more.

Later he layered in a subscription tier for teams too — and switching his payment stack to Lemon Squeezy reportedly drove a ~1300% MRR jump and $200k+ in a few months by making purchases and tax/global payments frictionless.

Speed wins the window. Structure — a product with no costs to scale — is what keeps the margins fat.

Move 3: Ship relentlessly, in public

Launch was the start, not the finish. Tony shipped updates constantly (feature after feature), which kept momentum and word-of-mouth alive. He'd cross 4,000+ paying users within 3 months.

And all of it rode on an audience he'd been building openly: his X following went from ~100 to 180,000+ by consistently posting his dev process, revenue numbers, and failure stories. That audience is why a 5-day build had thousands of eyeballs on day one.

The honest asterisk

The "shipped in days" headline is real — and incomplete. Tony had already quit his job ~2 years earlier, shipped several products (some that flopped), and grown a real audience. So the fast build sat on top of years of skill and a warm launchpad. The speed was genuine; the head start was too. Copy the moves, not the fantasy of zero groundwork.

How to steal this (for real)

1. Watch for platform moments. A new API, model, or marketplace = a short land-grab for the better wrapper. When one drops, ship in days, not months.
2. Engineer your costs out. Can users bring their own key/account/data so you don't carry the expensive part? BYOK, self-hosting, or usage passed through to the provider can turn a cost-heavy idea into a high-margin one.
3. Make the "yes" easy. A one-time price often converts better than another subscription — especially for developers and prosumers. Use a payment tool (like Lemon Squeezy/Stripe) that handles global tax and checkout so you don't lose sales to friction.
4. Grow the audience before you need it. Tony's 5-day launch worked because ~180k people had watched him build for years. Start posting your process now, at zero revenue — it's the launchpad for everything later.
5. Ship after you ship. The first version made money; the next hundred updates made a business. Momentum is a feature.

The window is open on something right now. The only question is whether you ship this week or "eventually."

What new platform or model is sitting there, right now, without a great tool wrapped around it?

---

Sources (founder-reported / secondary — verify before quoting): Tony's own newsletter — Making $22K in 7 days and the $500K milestone, Starter Story breakdown, Lemon Squeezy case study, and Tony's X (@tdinh_me). Revenue figures are founder-reported and change over time.