Introduction
In the summer of 2020, a Dutch freelancer working from Bali did something that looks obvious only in hindsight: he emailed the CTO of OpenAI and asked for access to GPT-3. Danny Postma got in with the very first batch. Within weeks he had wired the model into Headlime — a headline-writing tool he had built out of a $19 eBook of headline formulas — making it one of the first GPT-3 applications running in production anywhere in the world. Eight months later, he sold Headlime to Conversion.ai (now Jasper) for a reported $1 million.
Postma's story is the purest example of a particular indie-hacker skill: not predicting the wave, but already being in the water when it arrives. He didn't invent the AI copywriting category. He was simply the fastest person to bolt a frontier model onto a boring, well-understood job — writing headlines — and the first to sell the result. Then, instead of retiring on the exit, he did it again: ProfilePicture.AI, Deep Agency, and HeadshotPro, a Singapore-based AI photography company that says it has generated over $10 million in revenue since launch.
The Story
Before the exit, there were the unglamorous years. Postma was freelancing in conversion optimization and building side projects from Bali. His first notable one was Landingfolio, a WordPress gallery of landing-page designs — useful as a reference tool, weak as a business. The pattern that would define his career was already forming, though: find a sharp, specific job people already pay to have done, ship a thin product for it, and tell the truth in public while it grows.
The $19 eBook came out of that freelancing work. Postma had collected 200 headline formulas into an interactive book and was selling it for $19. It was not a hit. But instead of moving on, he asked the question that became Headlime: what if the book were software? The first version was almost embarrassingly simple — a copy spinner that swapped variables into his headline templates. Then GPT-3 arrived, and the spinner became an AI writer.
The access story matters because it reveals the method. Postma didn't wait for public API docs, tutorials, or case studies. He reached out directly to OpenAI's CTO Greg Brockman and got into the first access cohort. He has described being astounded by what the model could generate. Within weeks, Headlime was one of the first products in the world with GPT-3 in production, alongside the early Jasper and Copy.ai wave. Being early wasn't a branding exercise — it was the entire moat. For months, Headlime could do something almost no other tool on the market could do.
He also documented the whole run in public. Postma built Headlime openly on Twitter — sharing revenue numbers, experiments, and mistakes as they happened — and that transparency compounded into an audience that would follow him into every later project. By 2026 that audience had grown to roughly 180,000 followers, a distribution asset he never had to pay for and one that made each subsequent launch cheaper than the last. Build in public is often sold as a feel-good philosophy; in Postma's case it was a balance-sheet item.
Distribution was demand-first, not audience-first. Postma has said he used Ahrefs for keyword research before and during the build: find search terms with real demand and weak competition, then build the product that those searches are already looking for. He launched with a tiered lifetime deal on Product Hunt to buy initial traction and reviews — 80+ people joined the waitlist shortly after he announced it. Then a TechCrunch feature landed, and by his own account it "exploded" growth. The numbers tell the rest: roughly $1,000 in monthly recurring revenue in December 2020, about $20,000 by February 2021.
The exit was as fast as the build. Conversion.ai — the company that would rebrand as Jasper — acquired Headlime in early 2021, about eight months after launch, for a reported seven-figure sum of around $1 million. People who have written up the deal note that the acquisition itself took only one to two weeks from first conversation to close — the kind of speed that's only possible when a bootstrapped company has clean books, recurring revenue, and a single decision-maker.
What happened next is the part operators should study more carefully than the exit. Postma didn't become a full-time investor or a course seller. He built a portfolio — and he built it inside a single umbrella. Every project since Headlime has lived in the same neighborhood: AI applied to a concrete creative job. Headline formulas became copy generation; copy generation became profile pictures; profile pictures became professional headshots. Each launch reused the previous one's knowledge — prompting, inference costs, checkout flows, affiliate playbooks — so the second company was cheaper to build than the first, and the third cheaper still. Most founders start from zero every time. Postma started from Headlime every time. When Stable Diffusion was released, he shipped ProfilePicture.AI in about 30 hours — fast sales, then demand cooled as the novelty wore off, and he moved on without sentimentality. Deep Agency followed. Then, in March 2023, HeadshotPro: AI-generated professional headshots. Early write-ups cited roughly $100,000 in revenue in its first 14 days, and the company's own about page (September 2026) claims more than 17 million headshots generated, 200,000+ photoshoots, and over $10 million in revenue since launch. Along the way he bought adjacent AI products — including Hesti AI and Dreamchanted — and redirected their traffic into his own properties, a quiet roll-up strategy most solo founders never attempt. He now runs the portfolio under a studio called Postcrafts, with the operating company based in Singapore, and documents the process to an audience of around 180,000 followers on X.
Lessons
Danny Postma's run offers three mechanics worth stealing: 1. Be first on the platform shift — in week one, not year one: Postma didn't wait for GPT-3 tutorials; he emailed the CTO and shipped while the model was still in private beta. Copy it: when a major model or API opens up, give yourself a two-week deadline to ship the thinnest useful wrapper on top of it. The "first mover" window on a platform shift closes in weeks, and the founders inside it get distribution, press, and acquisition interest that latecomers can never buy back. 2. Turn inventory into products: Headlime was a $19 eBook before it was a SaaS company. The book wasn't a failure — it was R&D with a price tag. Audit everything you've already made: books, templates, checklists, spreadsheets, swipe files. Each one is a product candidate if software can do the same job ten times faster. Your next company may already be sitting in your downloads folder. 3. Build for demand that already exists: Postma ran Ahrefs keyword research before writing serious code — find the searches with real volume and weak competition, then become the answer. Pair it with an irresistible launch offer (his tiered lifetime deal on Product Hunt) to convert early attention into reviews and word of mouth. Don't build the product and then go hunting for demand; let the demand pick the product.
Closing
Eight months from eBook to exit. Danny Postma's edge was never the model — models get commoditized. His edge was being in the water before the wave arrived, with a product thin enough to ship in weeks and a distribution plan built on demand that already existed. The $19 eBook nobody remembers became the $1M company everybody cites. Timing is a strategy — but only if you ship fast enough to use it.