He Built a Better ChatGPT in a Day: a Solo Developer's $500K First Year, No Team, No Funding
Tony Dinh, a solo developer from Vietnam, sells TypingMind — a faster, nicer interface built on top of the ChatGPT API. With no team, no investors, and no funding, he registered the domain and built the first version in a single day, the very week OpenAI opened its API in March 2023; he made $22,000 in the first week and crossed $500,000 in the first year, now clearing as much as $83,000 in a single month at ~85% profit. It wasn't luck: he had spent years building small products in public on X (growing to 97,000 followers and selling Black Magic for $128,000 and Xnapper for $150,000) before he had anything big to sell. Startup cost was essentially $0 — his own code, a domain, and pay-as-you-go API.
Process
Tony Dinh — a solo developer from Vietnam — sells TypingMind, a faster, nicer interface built on top of the ChatGPT API. With no team, no investors, and no funding, he crossed $500,000 in revenue in his first year and now clears as much as $83,000 in a single month — from a startup cost of essentially $0.
Stage 1 — Seven years of paychecks, then a deliberate leap: quitting a $105K job with two years of runway
Tony Dinh was born in Vietnam in 1993 — no family money, no Silicon Valley network, no startup pedigree. From 2014 to early 2021 he worked his way through five startups and one big corporation as a developer, climbing to a $105,000-a-year salary. Then COVID hit, remote work became the norm, and Tony fell down the indie-hacker rabbit hole: he discovered Indie Hackers, got hooked on the podcast, and made the success stories of solo developers his daily dose of motivation. In 2021 he made the jump on purpose, not on a whim — his high-paying years had given him enough savings to survive about two years in Vietnam with zero income. That runway was the thing that let him take real risks without panicking.
Stage 2 — Years of small bets, built in public: an audience of 97,000 before the big win
He didn't strike gold on day one. When he quit, his early products Black Magic and DevUtils were together making just $600 in MRR. So he kept shipping small products and "building in public" on X (Twitter) — sharing his revenue, his failures, and his process honestly, long before he had anything impressive. Black Magic, a Twitter analytics tool, climbed to $14K in monthly recurring revenue and later sold for $128,000; Xnapper, a screenshot-beautifying tool, did about $4K a month and sold for $150,000. Each product taught him to ship, to find an audience, and to sell — and along the way his X following grew from near-zero to 97,000 in about two years. People trusted him because he'd been honest, consistently, when he had nothing to gain from it.
Stage 3 — The one-day head start: turning a brand-new API into a product the same day it shipped
On March 1, 2023, OpenAI opened the ChatGPT API to developers. Tony had been using ChatGPT obsessively and was frustrated by its clunky interface every single day — it was slow, it logged him out, and there was no good way to search his old conversations. The moment the API dropped, he saw the opening: he registered TypingMind.com that same day and built the first version in about a day, on a stack he knew cold (NextJS, TailwindCSS, Node.js) — a faster, nicer interface sitting on top of the API, with the search and speed ChatGPT lacked. He wasn't inventing a new AI; he was wrapping a powerful-but-clunky tool in a great experience. By the end of that first week, he'd made $22,000.
Stage 4 — One person, 171 updates, ~85% margins, a million-dollar year
He didn't stop at launch. In the first year alone he shipped 171 updates, listening to power users and steadily raising the price as the product got better — from a $9 one-time purchase to a $39 license, then premium tiers and subscriptions for cloud sync and teams. TypingMind crossed $500,000 in revenue in its first year, and by late 2024 — including enterprise and B2B deals — Tony was clearing as much as $83,000 in a single month at around 85% profit. All of it solo: no team, no ad budget, no funding, no investors. Startup cost was basically zero — his own code, a domain name, and pay-as-you-go API fees that scale only when customers do.
"Be honest, consistently, before you have anything impressive to show — and when a new platform opens, be first." — Tony Dinh (paraphrased from his public posts)
Source: Starter Story · Indie Hackers · Tony Dinh's public posts on X (@tdinh_me)
Thinking
Insight 1: When a platform-level door opens, the first week is everything
The day OpenAI released the ChatGPT API, Tony shipped. A brand-new platform capability is a narrow window where one person can beat everyone — because the giants are slow and the crowd hasn't arrived yet. Watch for the launch; move in days, not months.
Insight 2: Build in public — the audience and the trust arrive years before the big win
TypingMind sold from day one because Tony already had 97,000 people who trusted him — earned over years of honest posting, long before he had anything impressive. An audience is the asset that makes a launch instant; you build it quietly, before you need it.
Insight 3: You don't have to invent something new — a better experience on a giant's product is a huge opportunity
He didn't build a new AI. He built a better interface on top of ChatGPT — solving the daily friction he himself felt. Wrapping a clunky-but-powerful tool in a great experience is one of the most reliable software businesses there is.
Insight 4: A string of small "failures" and small exits is the road to the big one
Black Magic and Xnapper weren't detours — they were practice, savings, and trust-building that made TypingMind possible. The "overnight" win is almost always sitting on a pile of small bets nobody clapped for.
Insight 5: One person plus AI-era leverage equals software-level margins
171 updates, ~85% profit, no team. In the AI era, a single capable builder can run a business that used to need a whole company — the leverage is real, and it compounds.
Action
Step 1: Pick a tool you use every day whose experience frustrates you
The friction you feel daily is market research. Find a powerful tool with a clunky experience — the gap between "powerful" and "pleasant" is where a product lives.
Step 2: Watch for platform-level launches and move in the first week
New APIs, new models, new app stores — when a big new capability opens up, the first week is a gift. Have something rough live before the crowd shows up.
Step 3: Build in public on one platform, long before you have a product
Pick one platform (X, LinkedIn, YouTube) and share your work and your process honestly, consistently — for months. The trust you bank now is what makes your future launch sell on day one.
Step 4: Ship a minimal version fast and charge from day one
Don't wait for perfect. Build the smallest useful version and put a price on it immediately (one-time or subscription). Getting paid on day one tells you it's real.
Step 5: Iterate relentlessly and stack products
Listen to users and ship constantly (Tony shipped 171 updates in a year), then build more products around the same audience. Momentum and catalog both compound.
Not for you if: you won't learn to build software (this path needs code — though AI is lowering that bar fast); you want truly passive income (it takes relentless iteration and years of building in public); or you want money today (the "overnight" win sat on top of years of small bets).