Latest / The Indie Hacker Podcast with Fexingo: Solo Developers, SaaS Side Projects, and Independent Tech / How a Solo Dev Hit 10K MRR With a Product That Sells AI Prompts
Transcript
- Lucas: So there's this solo developer, goes by the handle PromptSmith online — real name is Alex Chen — who in late 2024 launched a simple marketplace for AI prompts. Not a new idea on the surface, but he hit ten thousand dollars in monthly recurring revenue within eight months, and he did it completely bootstrapped, no outside funding. Luna: A marketplace for prompts — like, selling prompts for ChatGPT, Midjourney, that kind of thing? Lucas: Exactly. Think of it as an Etsy for prompts. Creators upload prompt templates for image generation, code writing, copywriting, whatever. Buyers pay a small fee — typically two to five dollars per prompt — and the platform takes a thirty percent cut. Alex built the whole thing with a no-code tool called Bubble and a Stripe integration. Luna: Right, so the tech stack is minimal. But the challenge with any two-sided marketplace is getting both sides to show up. How did he solve the cold-start problem? Lucas: He did something pretty clever. Instead of trying to recruit sellers first, he created fifty prompts himself using GPT-4 and Midjourney, listed them on the platform, and then started a niche blog about AI prompt engineering. He wrote posts like 'Ten Prompts to Generate Consistent Characters in Midjourney' and 'How to Write SQL Queries with Natural Language.' Luna: So he was essentially acting as the entire supply side while building demand through content marketing. That's a lot of hats. Lucas: It is, but he kept his costs near zero. The blog cost him about fifty dollars a year for hosting, and Bubble cost eighty dollars a month. His only real expense was time — about three hours a day after his full-time job. Within the first month, he'd made four hundred dollars from prompt sales, mostly from organic traffic. Luna: And that initial traction gave him credibility to start onboarding other prompt creators? Lucas: Exactly. Once he had a few dozen sales and some five-star reviews, he reached out to prompt creators on Twitter and Reddit — people who were already sharing free prompts — and offered them a revenue share. He'd handle hosting, payments, and marketing; they just had to upload prompts and update them occasionally. Luna: What was his split? Lucas: Seventy-thirty in favor of the creator, which is actually generous for a marketplace. The standard for something like Etsy or Gumroad is closer to eighty-twenty for the creator. But Alex figured that a lower take rate would attract higher-quality creators, and he was right. Within three months, he had two hundred creators and over a thousand prompts. Luna: But more supply doesn't automatically mean more demand. How did he scale the buyer side? Lucas: This is where the affiliate program came in. Alex set up a system where anyone with a blog or YouTube channel could promote specific prompts and earn a twenty percent commission on sales. He gave affiliates unique discount codes to track conversions. Within a month, he had about fifty active affiliates, mostly small tech bloggers and AI tutorial creators. Luna: So his marketing cost was essentially variable — only paid out when a sale happened. That's smart for a bootstrapper. Lucas: Exactly. And the affiliate model worked especially well because AI prompts are a 'high-intent' product. Someone watching a tutorial on how to generate better images is already primed to buy a prompt. The conversion rates were around five to eight percent, which is excellent for a five-dollar digital product. Luna: I'm curious about the economics at scale. If the average prompt sells for three dollars and the platform takes thirty percent, that's ninety cents per sale. To hit ten thousand MRR, you'd need about eleven thousand sales a month. That's a lot of transactions. Lucas: His average sale is actually higher — around four fifty — because many buyers purchase bundles of five or ten prompts. And about forty percent of revenue comes from a subscription tier. For nine ninety-nine a month, buyers get unlimited access to a curated library of one hundred prompts, which are updated weekly. That subscription alone accounts for about four thousand dollars of his MRR. Luna: Ah, so the subscription is the real revenue driver. That makes the marketplace more of a customer acquisition funnel. Lucas: Exactly. The one-off purchases are basically a trial. Once people buy a few prompts and see the value, many upgrade to the subscription. Alex told me the conversion rate from one-time buyer to subscriber is around fifteen percent, which is pretty solid for a digital product. Luna: So what's the biggest challenge he's facing now? Quality control? I imagine when you have hundreds of creators, some prompts are going to be duds. Lucas: That's exactly the issue. Alex has had to implement a curation system. Every prompt submitted goes through a two-step review: first an automated check to make sure it's not a duplicate or a direct copy of something from Reddit, then a manual test by Alex or one of his three part-time contractors. They run the prompt in the relevant AI tool and rate the output. Only prompts that score at least four out of five get published. Luna: That's a lot of overhead for a solo founder. How many prompts are they reviewing per week? Lucas: About two hundred. And the rejection rate is roughly thirty percent. Alex says it's the part of the business he'd most like to automate, but he hasn't found a reliable way to automatically assess prompt quality. He's experimented with using AI to judge AI prompts, but the results were too inconsistent. Luna: Makes sense. There's a certain human judgment involved in whether a creative output is actually good. So with ten thousand MRR, is he full-time on this now? Lucas: He quit his job as a backend developer about two months ago, when MRR hit seven thousand. He says his take-home after platform fees, affiliate commissions, and contractor payments is about sixty percent, so around six thousand dollars a month. That's enough to live on in Austin, Texas, where he's based, though he's reinvesting most of it back into content marketing and affiliate outreach. Luna: Speaking of content marketing — that blog you mentioned earlier. Is it still a big driver? Lucas: It's actually the primary acquisition channel. The blog now gets about fifty thousand monthly visitors, almost all from organic search. Alex writes two to three posts per week, mostly tutorials that naturally mention prompts from his marketplace. He also does keyword research targeting long-tail queries like 'Midjourney prompt for anime style' or 'ChatGPT prompt for resume writing.' Luna: So he's really doubling down on SEO. That seems like a sustainable moat — content that ranks takes time to build and competitors can't easily replicate overnight. Lucas: Right. And he's also started a YouTube channel where he screen-records himself building prompts. It's surprisingly popular — he has about twelve thousand subscribers now. The video format works well because people can see the exact prompt and the output side by side. Luna: It's interesting that he's building all these assets — blog, YouTube, marketplace — essentially as a solo operation. How does he manage the cognitive load? Lucas: He's very systematic. He blocks out his mornings for content creation — writing, recording, editing — and afternoons for marketplace operations: reviewing prompts, answering support emails, updating the subscription library. Evenings are for experimenting with new AI tools and finding trends. He told me he uses a tool called Todoist with a pretty rigorous priority tagging system. Luna: If today's episode gave you some practical ideas for your own side project, we want to mention that the show stays ad-free thanks to listener support. You can buy us a coffee at buy me a coffee dot com slash fexingo. It's a small way to keep these conversations going without interrupting them. Lucas: Yeah, we really appreciate everyone who chips in. It lets us focus on finding stories like Alex's without worrying about sponsorships. Now, one thing I wanted to ask you, Luna — do you think this kind of AI prompts marketplace has long-term staying power, or is it riding a wave that'll crest? Luna: That's the million-dollar question. On one hand, the underlying AI tools are improving rapidly, which could reduce the need for hand-crafted prompts. On the other hand, as AI becomes more ubiquitous, the demand for specialized, curated prompts for niche use cases might actually grow. Think about it — a lawyer using AI to draft contracts might pay for a set of prompts tailored to their jurisdiction. Lucas: That's a good point. Alex is already seeing that shift. His best-selling category is actually 'business productivity' — prompts for writing emails, creating presentations, analyzing data. And he's started partnering with small agencies to create custom prompt libraries for their clients. That could be a higher-margin B2B play. Luna: So the future might be less about a public marketplace and more about white-label prompt sets for specific industries. That's a natural evolution for a business that starts as a consumer marketplace. Lucas: Exactly. And Alex is already building toward that. He's got a waitlist for a platform where companies can create private prompt libraries for their employees. It'll be a monthly subscription per seat. If that takes off, his MRR could easily double in the next year. Luna: It reminds me of the early days of the app store — first it was all about consumer apps, then enterprise mobile solutions became huge. The same pattern might play out for AI prompts. Lucas: I think so. The key for Alex will be to not get too attached to the current business model. The marketplace might be the entry point, but the real value could be in the enterprise tools he builds on top of it. Luna: Any advice for listeners who want to replicate his success? What's the one actionable takeaway? Lucas: I'd say: start with a very specific niche. Alex didn't try to be the Amazon of prompts. He started with Midjourney image prompts, which had a passionate community and clear use case. Once he dominated that, he expanded to ChatGPT and other tools. Find the smallest viable audience first, serve them incredibly well, and then grow from there. Luna: That's solid. And don't underestimate the power of doing the work yourself at the beginning — being your own best creator and marketer. Lucas: Absolutely. Alex's willingness to create fifty prompts himself and write that first batch of blog posts was the foundation. No one else was going to do it for him.