Twitter killed his $14K/month app. He rebuilt — alone, from Vietnam, in 5 days

Read time: ~6 minutes.

In early 2023, Tony Dinh's most successful product was weeks from becoming unprofitable overnight. What he did next — alone, from Vietnam — is one of the best-documented builder stories in Asia, because he wrote it all down as it happened.

🎬 The Story

Tony Dinh is a software engineer from Vietnam who left full-time employment in 2021, after about seven years of writing code for other people, to build small products of his own. According to Starter Story's breakdown of his journey, those early products included DevUtils, a Mac toolbox for developers that went on to earn roughly $5,000 a month, and Black Magic, a set of analytics and CRM tools layered on top of Twitter.

Black Magic was the one that worked. Over two years he grew it from zero to about $14,000 in monthly recurring revenue — enough, by his own accounting, to make him independent. At one point he turned down a $500,000 acquisition offer for it.

Then the platform moved. Twitter's new API regime priced enterprise access at $42,000 a month, more than the entire business earned. In his newsletter, Dinh laid out the arithmetic plainly: "I currently only make ~$14K per month on average from the business." With API access set to be cut within weeks, he sold Black Magic to Hypefury for $128,000 — less than one year of its own revenue, and about a quarter of the offer he had declined months before. Twitter later added a $5,000-per-month tier, prompting one of the more human lines in his write-up: he "would have totally paid $5K/month if this had been announced one month earlier." He described his feelings about the sale, simply, as mixed.

Here is the part that makes the story worth studying rather than just sympathizing with. While Black Magic was dying, Dinh was already building. On March 1, 2023, OpenAI released the ChatGPT API. Dinh started coding a better chat interface on March 2, launched it on Twitter on March 6, and put it on Product Hunt on March 11, where it finished #1 Product of the Day. The product, TypingMind, was a static web app with no backend, no server, and no database — users bring their own API keys. By his own day-by-day account, revenue went $1K, $2K, $4K, $10K — and $22,000 total within the first seven days.

"I worked on the project like crazy. Added new features, improvements, and bug fixes every single day, multiple times a day."
— Tony Dinh, on TypingMind's launch weeks

The pace held. In his one-year reflection, Dinh counted 171 product updates in twelve months and $500,000 in cumulative revenue by February 2024, built on a hybrid of one-time licenses and a growing subscription tier. He also made a quieter, harder decision that year: he sold Xnapper, his screenshot tool, for $150,000 — per Starter Story — because splitting focus meant doing neither product justice. Two exits, neither triumphant, both deliberate.

As of his October 2025 update — his most recent detailed public accounting — TypingMind was earning roughly $130,000–160,000 a month, comfortably past $1 million a year, with the B2B Team version now more than half of revenue. Starter Story's profile lists the operation, at the time of its writing, as one person.

The takeaway

• Platform risk is a when, not an if. Black Magic died by someone else's pricing memo. TypingMind's bring-your-own-API-key design means no single vendor can repeat the trick.
• Speed can be the moat, briefly. Dinh shipped five days after the ChatGPT API launched, then defended the head start with 171 updates in a year. As he put it: being early to the niche is a huge advantage.
• Small exits are a tool, not a failure. $128K for Black Magic and $150K for Xnapper weren't jackpots — they were how a solo founder bought back his focus.

🌏 Around the Region

• Singapore — Bannerbear's open books. Jon Yongfook has documented his image-automation API from $859 MRR in May 2020 to $45K MRR and a team of seven by 2022, including his reasons for refusing venture funding. One of the region's best public build-logs.
• India — even giants have down years. Zerodha, the famously bootstrapped brokerage, saw revenue and profit fall roughly 15% in FY25 after regulatory changes, with founder Nithin Kamath saying "the time has finally come for business to pivot." Real stories include the down years.
• Korea → Palo Alto — the documentary channel that built itself. EO, Taeyong Kim's founder-documentary studio, has written up its own climb to 600K YouTube subscribers — 17 months for the first 200K, 8 for the last — including the cold-call years nobody filmed. Its camera grammar is a house influence here.

Know a builder who deserves the spotlight? Tell us — just reply to this email. We read everything.

— The Editorial Team