The canonical 'wrap a model, charge monthly' play. One person, no employees, no VC — train on a user's selfies, sell AI photoshoots. Sister product Interior AI reportedly runs >99% margin after the GPU bill.
Not 'AI side hustle' clickbait — the four real money loops (products, services, content, building) with the operators and numbers behind each.
"Using AI to make money" is the search behind a thousand grifty videos — but a handful of solo operators have made it boringly real. Pieter Levels (@levelsio) runs PhotoAI at roughly $138K/mo and a ~$200K/mo portfolio with zero employees. Danny Postma's HeadshotPro crossed ~$300K/mo selling AI headshots. Two teenagers built Cal AI — a photo-calorie app — to ~$30M/yr and sold it to MyFitnessPal. Maor Shlomo bootstrapped Base44 to ~$200K/mo in six months and sold to Wix for $80M cash. None of these were 'prompt packs.' They are four distinct money loops — sell an AI product, sell AI services, monetize AI content, or build-and-sell — and which one fits you depends on whether you can ship code, sell, or make content. This page maps the real version of each.
The canonical 'wrap a model, charge monthly' play. One person, no employees, no VC — train on a user's selfies, sell AI photoshoots. Sister product Interior AI reportedly runs >99% margin after the GPU bill.
AI corporate headshots from casual selfies — sells into a B2B/team use case (LinkedIn, company directories) with a higher willingness to pay than consumer photo apps. Distribution via SEO + maker audience.
Built by two teenagers — proof a photo-to-value AI wrapper (snap your meal, get calories) can hit real scale fast via TikTok-native distribution, then exit to an incumbent hungry for AI.
A vibe-coding app builder, but more importantly a clinic in build-and-sell: one founder, no funding, sub-Y exit window. Both the tool you can build on and the playbook to copy.
The vibe-coding rail letting non-engineers ship sellable apps from prompts — 100K+ projects/day. The 'I can't code' on-ramp to the sell-a-product and build-and-sell loops.
A SaaS boilerplate sold to other makers — the 'sell shovels to AI builders' angle. Lou's whole portfolio (CodeFast, DataFast) shows how one solo maker stacks multiple AI-era income streams off an audience.
The lowest-friction storefront for the build-and-sell-a-digital-product loop: list an AI-made product, sell to 3 people, validate demand before building anything bigger. Pairs with the content loop for distribution.
The rail behind the AI-content money loop — auto-cuts long video into viral shorts. The tool faceless/clip operators use to turn one recording into the volume that ad-revenue and sponsorship income require.
These are not theoretical: PhotoAI ~$138K/mo with zero employees, Cal AI ~$30M/yr, Base44 an $80M exit in six months. A one-person team really can reach numbers that used to require a funded company — the AI wrapper compresses the team you'd otherwise need.
An AI product can run absurd margins once the model bill is covered (Levels claims Interior AI >99%). And you can validate on Gumroad or a landing page in a weekend — sell to 3 people before you build the real thing, the way every durable solo maker does.
You don't need to code. A seller does the services loop (AI automation agency); a maker does products via Lovable/Base44; a creator does the content loop with OpusClip; a coder does build-and-sell. Pick the loop your existing skill already feeds.
The loudest voices sell courses about using AI to make money, not the loop itself. Treat YouTube/X gurus as marketing, not benchmarks — model yourself on operators with public, verifiable revenue (Levels, Postma, Lou), and assume any 'passive $10K/mo' promise is a funnel.
A bare 'ChatGPT wrapper' is a weekend build for the next person and a feature away from being eaten by the model provider. Defensibility lives in a sharp niche, distribution, and brand — not the prompt. Undifferentiated tools race to zero fast.
In early 2026 YouTube suspended thousands of mass-generated AI channels under its inauthentic-content policy; only ~3% of automation channels ever monetize. App stores, ad networks, and model T&Cs can all change the rules overnight. Don't build your whole income on one platform's mood.
Maker or vibe-coder who can ship a narrow, sharp tool
Seller comfortable with outreach and client work, code optional
Coder / lone engineer who'd rather flip product than serve clients
Every one of these loops converts internet attention into revenue with no team and no inventory — exactly the motion this archetype runs. Distribution is the moat, and a net-savvy-solo already lives in the channels (X, YouTube, SEO) where AI products and services get sold.
The highest-ceiling loops — sell-a-product and build-and-sell — reward someone who can ship a tight app and iterate fast. PhotoAI, HeadshotPro and Base44 are all one-builder products; this is the archetype's home turf, now with AI compressing the build.
The services loop and a Gumroad digital product both start nights-and-weekends with near-zero capital and pay per outcome, stacking cleanly on a day job until the income justifies going full-time — the stacker's exact entry pattern into AI money-making.
5 min · 12 questions · Free · Get your archetype + top 3 matching tracks
Take the quiz →