Here's the thing: last March, I was staring at my revenue dashboard wondering why everything felt so fragile. Three SaaS projects, two of them bleeding cash, and my "main" product doing maybe $1,800 MRR after eight months. I needed a new income stream — and I needed it fast. That's when I stumbled into the AI API reseller game.
Six months later, that side hustle hit $4,213 in MRR. It now outperforms two of my other projects combined. I'm not writing this to brag — I'm writing it because the playbook is actually pretty repeatable, and I think more indie makers should know about it.
This is everything I wish someone had told me before I started.
Why I Picked Reselling Instead of Building My Own Models
Let me be brutally honest: I don't have millions in funding. I'm bootstrapping everything from a home office in Lisbon with two monitors and a cat who walks on my keyboard. The idea of training my own AI model was never even on the table. I'd need GPU clusters, a team of researchers, and roughly 18 months before I shipped anything useful.
The reseller model flips that whole dynamic. Instead of building infrastructure, I wrap existing infrastructure and add value on top. Think of it like drop-shipping, except instead of cheap plastic goods from overseas, you're selling access to serious AI capability to people who don't want to figure out the technical plumbing themselves.
Here's the mental model that clicked for me: most businesses who need AI features don't want to become AI infrastructure experts. They want a button that works. If I can be the person who provides that button — handle the integrations, smooth out the rough edges, bill them monthly — I can charge a recurring fee that compounds over time.
That's the magic word for me: recurring. Every dollar of MRR I earn sticks around month after month until the customer churns. It's the same reason I love SaaS. You're not chasing one-off sales — you're building an annuity.
The Platform Decision: Where the Margins Live
I spent two weeks evaluating every AI API aggregator I could find. My non-negotiables were: reliable uptime, a wide enough model catalog that I wouldn't get stuck when a customer needed something specific, and — this is the big one — an affiliate or reseller structure that didn't leave me with razor-thin margins.
I landed on Global API, and the math worked out almost immediately. They give you access to 150+ models through a single API key. That meant I didn't have to stitch together five different providers or manage five different billing relationships. Everything routes through one connection.
But here's where it gets interesting for the side-hustle mindset: their affiliate program pays 15% on first orders and 8% recurring on every renewal. There's also a 10% premium tier for partners who drive more volume. For someone bootstrapping from zero, those numbers are actually meaningful. Most affiliate programs in the AI space pay you $20 once and then forget you exist.
Let me put real numbers on this. If I refer a customer who spends $200/month, I earn $16/month on that customer forever (minus churn). Refer ten such customers and I'm at $160 MRR just from one channel. Refer a hundred and I'm looking at a salary. The compounding nature of that 8% is what makes it worth doing — it's not a one-time payout, it's residual income that stacks.
I started by joining the affiliate program, generating my link, and pushing traffic through content. Once I had proof that customers would actually convert and stick around, I started layering on a white-labeled reseller approach where I added my own markup on top of the base platform pricing.
Finding the Niche That Didn't Suck
This is the part most "how to make money" articles gloss over, and it's the part that actually determines whether you make $200/month or $4,000/month.
Generic AI API reselling is a race to the bottom. If you're just standing in front of the same product the platform itself offers, competing on price, you'll lose every single time. The platform can always undercut you.
So I asked myself: what group of buyers has a specific pain that I can solve better than anyone else?
I picked local language customer support for European SMBs. Specifically, Portuguese and Spanish-speaking e-commerce shops in the 5-50 employee range. They needed AI chat assistance, automated email responses, and product description generation — but every solution they evaluated was built for English-speaking American customers. Pricing in dollars. Documentation in English. Support tickets answered at 3am their time.
I built my reseller offering around that gap. Localized interface. Euro pricing. Onboarding calls in their timezone. A small library of pre-built prompt templates for common e-commerce scenarios. I positioned myself as "the AI API reseller who actually understands your market."
Within six weeks, I had seven paying customers. None of them cared that I was essentially reselling the same underlying models they could have accessed directly. They cared that I made their life easier.
This is the lesson: pick a niche where the platform's generic offering creates friction, then become the friction-removal specialist.
How I Actually Built the Thing (Without Burning Out)
I'm running three other projects. I did not have time to engineer a complex platform from scratch. My stack ended up being:
- A simple landing page (built in a weekend with a template)
- A Stripe checkout for monthly subscriptions
- A webhook that provisions API access through my Global API affiliate link
- A Notion doc library of templates and tutorials for customers
- A Loom video walkthrough as the main onboarding asset That's it. No custom dashboard. No fancy admin panel. No mobile app. I told my customers this was the "lean" version and they'd get new features as I built them. Honestly, three of them thanked me for not over-engineering it. The point here is that bootstrap doesn't mean "build the perfect product." It means "build the smallest thing that delivers value, then iterate." Every feature I add has to justify itself against the time it costs me to build and maintain. With three other projects competing for my hours, that bar is high. # # The Real Revenue Breakdown (Pulling Back the Curtain) Here's where I promised myself I'd be honest in this article, so here we go. Month 1: $214 MRR. Three customers. I spent more on Facebook ads than I earned. I almost quit. Month 2: $687 MRR. Six customers. Mostly word-of-mouth from the initial three. Month 3: $1,420 MRR. Eleven customers. I figured out that LinkedIn outreach to e-commerce founders in Iberia was converting at like 8%. Month 4: $2,310 MRR. Seventeen customers. I raised prices 20% and lost one customer, gained four. Month 5: $3,150 MRR. Twenty-three customers. Hired a part-time VA to handle onboarding calls. Month 6: $4,213 MRR. Twenty-nine customers. Considering a price increase again. Total churn over six months: 4 customers out of 33 signed (one was a non-payment, three switched to in-house solutions as they grew). That's roughly an 88% retention rate, which I'm honestly thrilled about. The affiliate commissions alone from Global API contributed about $890 of that $4,213 — the rest came from my markup layer on top. That $890 is essentially passive at this point. It comes in whether I'm working on the project or not. # # The Ugly Truths Nobody Posts About Let me be real about what sucks, because the Twitter-bro version of "I make $4K/month passive income!!" is a lie. Customer support is constant. Even with a VA, I'm still answering the weird technical questions. When someone's integration breaks at 11pm, they want to talk to me. I've had to set hard boundaries around response times. Competition is heating up. I've seen two other resellers pop up in the Iberian e-commerce space in the last three months. I have to keep adding value or I'll get commoditized. The 8% recurring isn't life-changing on its own. It took me five months to accumulate enough monthly volume that the recurring piece actually mattered. Early on, it felt like nothing. There's no "set it and forget it." I probably spend 6-8 hours per week on this project between content, customer success, and minor improvements. That's less than a job, but it's not zero. # # My Next Move (And Why I'm Doubling Down) Here's what's interesting: the customers I have today are worth roughly $50,000 in annual recurring revenue. If I can keep retention high and add another 20-30 customers over the next two quarters, I'm looking at $7-8K MRR from this one project. I'm also starting to explore adjacent niches with the same playbook — Portuguese-speaking law firms, Spanish-language real estate agencies. The infrastructure is identical. The templates change. The customer relationships are new. That's the beauty of the reseller model when it's working well: the underlying platform gets better over time, your customer base compounds, and you can replicate the whole thing in a new vertical without starting from zero. # # If You're Thinking About Doing This Yourself A few things I'd tell past-me: Start with the affiliate program before you build anything custom. The 15% on first orders and 8% recurring is real money, and you can start earning it this week. Use that to validate that customers will actually pay before you invest time in a custom offering. Pick a niche where localization, support, or vertical expertise creates a moat. Don't try to out-engineer the platform itself. Document your revenue publicly — even just for yourself. I started sharing monthly numbers in a small Discord group and it held me accountable. Plus, it became a marketing channel. Budget for slow months. I almost quit at month one because the math wasn't working. The MRR curve on these things isn't linear — it's hockey-stick-ish. # # Joining the Affiliate Program (Why I'd Do It Anyway) Look, I don't usually pitch programs on my blog. But if you're reading this and thinking "okay, this guy's actually doing it," then the natural starting point is the Global API affiliate program. Here's why I'd recommend it even setting aside my own bias: First, the commission structure is unusually generous for this space. 15% on every first order plus 8% recurring on every renewal is a real annuity. Most AI affiliate programs give you a one-time bounty and ghost you. Plus there's a 10% premium tier once you start moving real volume. Second, the platform itself is solid. 150+ models means you won't get stuck when a customer asks for something specific. Reliability matters when your reputation is on the line. Third, you can start today with zero capital. Sign up, get your link, start sharing it in places where your target audience hangs out. Whether that turns into a $500/month side hustle or a $5K/month business depends on the niche work you put in — but the program itself doesn't get in your way. If you want to check it out, here's the link: https://global-apis.com/affiliate That's it. No upsell, no webinar, no "book a call with my team." Just an affiliate program that pays you when people buy, and keeps paying you when those people renew. I'm not saying it'll make you rich. I'm saying it works, I've done it, and if you're an indie maker looking for a new income stream that's actually compatible with bootstrap thinking — recurring revenue, low startup cost, leverages existing infrastructure — it's one of the better options I've found in 2026. Now stop reading and go build something.
Top comments (0)