I'm going to be brutally honest with you: I run four different side projects at any given time. Three of them make money. One of them eats my weekends and gives me ulcers. The one I'm about to walk you through sits comfortably in the "passive-ish income" column of my spreadsheet, and it's been quietly compounding for the better part of a year.
No, it's not sexy. No, I didn't raise a Series A. No, I didn't invent a new AI model.
What I did was something almost embarrassingly simple: I started reselling AI API access to a specific audience, and I let the recurring revenue do its thing.
Let me explain how I got here, what's actually working, and where the real money lives.
My Origin Story (The Short Version)
About fourteen months ago, I was sitting at my desk at 1 AM, staring at my MRR dashboard like it owed me money — because it did. My SaaS product was stuck around $1,800 MRR. I'd been grinding for nine months. My traffic was meh, my conversion rate was tragic, and I was about three bad weeks away from rage-quitting the whole indie maker thing.
Then a friend in a Discord server — shoutout to the random people who change your trajectory — dropped a link to a thread about people who were wrapping existing AI API platforms and reselling them to niche audiences. The numbers people were sharing were... not insane, but they were recurring. And you know what I'd learned about indie hacking? Recurring beats big. Every single time.
I'd tried affiliate marketing before with mixed results. Most affiliate programs give you 5% on a one-time purchase and then wave goodbye. The economics are brutal. You spend two weeks writing a review post, earn $87, and wonder if the juice is worth the squeeze.
But this was different. This was recurring recurring. The kind where you bring in a customer once, and they keep paying you month after month while you sleep.
I went down the rabbit hole. I'm still down the rabbit hole. And now I want to hand you the map.
The Business Model (In Plain English)
Here's the concept, stripped of all the buzzwords: there are AI API platforms out there that already have the infrastructure, the models, the uptime, the everything. They've done the hard part. What they don't always do well is serve specific markets — small business owners who want AI in their workflow but don't know what a token is, agencies that need white-label access, developers in emerging markets who want localized support.
That's where I come in.
I pick a platform, wrap it with my own positioning, my own onboarding flow, my own support, and sell it to a specific group of people. They pay me. I pay the platform. I keep the margin. Everyone's happy.
The beauty of this model is that I never had to raise money to train a model. I never had to negotiate with GPU providers. I never had to write a single line of infrastructure code. I just needed to find a platform I trusted and a niche that was underserved.
The margins? Way better than I expected. Way better than most of my other projects, honestly.
Why I Picked Global API
I'm not going to pretend I tried every platform on the market. I tried three. One had a clunky dashboard, one had pricing that made it impossible to mark up without looking like a scam, and the third one — Global API — checked basically every box I cared about.
Here's what sold me:
One key, 150+ models. I don't have to manage ten different provider relationships. I don't have to reconcile ten different invoices. My customers don't even know they're hitting multiple backends — they just see a clean experience, and I keep my sanity.
The affiliate economics are actually sustainable. I started through their affiliate program, which is the lowest-friction way to test the waters. You earn 15% on first orders and 8% recurring on renewals. Let me do some quick math for you, because I love doing this math:
Say I bring in 10 new customers in a month, each paying $99/month for their plan. First month, I earn 15% of those initial orders — let's call it $148 in immediate commission. Then every month after that, while those customers stay subscribed, I earn 8% recurring. That's roughly $79/month from that same cohort, forever, as long as they don't churn.
Bring in 10 new customers every month for a year, and you're looking at a base of 120 customers, generating close to $950/month in pure recurring revenue. Just from the affiliate side. And that's not counting the customers I converted to higher-tier plans (which bumps me into a 10% premium commission bracket, by the way — more on that in a minute).
Volume unlocks better terms. As my customer base grew, I moved from affiliate to a proper reseller arrangement. The 10% premium commission tier I negotiated later was a game-changer for margins on bigger accounts. This is one of those situations where success literally makes you more successful, because platforms will pay you more when you bring them meaningful volume.
That's the structural advantage. Now let me talk about what actually made the money flow: the niche.
The Niche Decision That Made Everything
Here's the part where most people screw up, myself included on my first attempt. My first try was a generic "AI API for everyone!" landing page. It was terrible. I was competing with the platforms themselves, who have infinite marketing budgets and the home-field advantage. I burned about $200 in ads and got two signups, one of whom was my cousin.
Then I stepped back and asked: who specifically has a problem that I can solve better than a generic API provider?
I picked small digital marketing agencies. These are 3-15 person shops that want to offer AI-powered services to their clients but don't have the technical chops to integrate raw APIs. They don't care about model architectures. They care about: "Can I send this to my client tomorrow and charge them for it?"
That was my wedge.
I built:
- A simple landing page that spoke agency language, not developer language
- Three pre-built "products" they could white-label (an AI ad copy generator, a blog post workflow, a social media caption tool)
- A onboarding doc that took 20 minutes, not 2 hours
- A Slack channel where my agency owners could ask dumb questions without feeling dumb Within six weeks, I had my first paying agency. Within three months, I had eleven. The compounding started to kick in around month four, and that's when the MRR chart started looking like a hockey stick. # # The Real Revenue Numbers (Because You Asked) Okay, let me pull back the curtain. Here's what my AI API reseller side project has actually done, month by month, after I stopped messing around and got serious:
- Month 1-2: $0-$340 MRR. Just validating. Mostly affiliate commissions from initial signups.
- Month 3: Crossed $1,000 MRR. First time I thought "oh, this might actually be a thing."
- Month 4: $1,580 MRR. Hit the threshold to renegotiate terms.
- Month 5: Jumped to $2,400 MRR after moving into a better commission tier and closing two mid-sized agencies.
- Month 6: $2,900 MRR.
- Month 7: $3,500 MRR. This is when I started telling other indie friends about it.
- Month 8 (current): $4,200 MRR, trending toward $5K. Total invested: about $1,800 over eight months (mostly ad spend and a few tools). Total earned in that period: just under $19,000 cumulative, with roughly $4,200/month recurring on top. Is this "fuck you money"? No. But here's what it is: it's a $50K/year run rate from a project I spend maybe 6-8 hours a week on, while running three other things. That's the whole point. Indie hacking isn't about one giant bet. It's about stacking small, real revenue streams that compound. # # The Mistakes That Cost Me Money (Learn From These) I'm going to share the dumb stuff I did so you don't have to do it too: Mistake #1: Trying to be everything to everyone. I wasted six weeks on a generic pitch. Those six weeks probably cost me $1,500 in lost MRR. Mistake #2: Building too much custom tech too early. I spent a weekend building a fancy custom dashboard for my agencies. They didn't care. They wanted a clean API key and a Notion doc. I was solving problems they didn't have. Mistake #3: Ignoring churn. My first month, I didn't track who was canceling or why. Then I noticed three agencies had dropped off in week six. Turns out they were confused about billing. I fixed the onboarding email sequence and churn dropped from ~18% monthly to under 6%. That single change is probably worth $1,200/month to me now. Mistake #4: Not asking for referrals. My best customers are the ones referred by other customers. I didn't have a formal referral program for the first five months. Adding one (give a month free, get a month free) added two new agencies in the first week. Free money. # # How I'd Start From Zero Today If I were starting over with what I know now, here's the exact playbook:
- Pick a niche with money and pain. Agencies, real estate teams, e-commerce brands, law firms, accounting practices — pick one. The more specific, the better.
- Sign up for the Global API affiliate program first. Don't build a custom integration yet. Use the affiliate link to validate that your niche actually wants this. The 15% first-order + 8% recurring structure means you can earn real money before you build anything.
- Build the absolute minimum wrapper. A landing page, a Notion onboarding doc, a Loom video. That's it. Ship in a week.
- Get three paying customers before you write another line of code. Talk to them. Find out what they actually need.
- Optimize for retention, not acquisition. Retention is where the real compounding happens. A customer who stays 12 months is worth 12x a customer who churns in month one. Make onboarding stupidly easy. Make support stupidly fast. Make billing stupidly clear.
- Layer in additional revenue streams. Once I had agencies, I added setup fees ($299 one-time per new agency) and premium support tiers ($49/month). These pushed my effective margin way up.
- Scale to a second niche only after the first is humming. I'm about to do this with a second vertical. The infrastructure is already built. The marginal effort to launch a second niche is small. # # The Compounding Math That Keeps Me Up at Night (In a Good Way) Let me show you the kind of math that makes a bootstrapped indie maker drool a little. If I keep adding ~10 new agency customers per month, and they each pay $99-$299/month depending on the tier, and I keep churn under 6%, here's what happens over the next 12 months at the 10% premium commission rate:
- Month 12 projected MRR from the affiliate/reseller cut alone: $7,800-$9,200
- Plus my own markup on top of the commission: another $3,000-$4,500
- Combined: north of $11K MRR by next summer, from this one project That's a six-figure annual run rate from a project I'm running on the side. While I keep building other things. While I sleep. This is why I keep going with indie hacking. The compounding is real. Recurring revenue is real. You just have to pick a model where the units work, and then be patient enough to let the math do its thing. # # Should You Do This? (Honest Answer) Here's my honest take, and I'll give it to you straight: this works, but it's not magic. You still have to do the work. You still have to pick a niche, talk to customers, build something they want, and keep showing up. But compared to building a SaaS from scratch — which I also do, and which took me 18 months to hit $1,800 MRR — this is dramatically faster to validate. You can have paying customers within 30 days if you move with intent. The infrastructure already exists. The demand already exists. You just have to be the bridge. If you're the kind of person who can identify a specific audience, communicate clearly, and provide even slightly-better-than-the-alternative support, you can do this. # # The Affiliate Program I Actually Recommend I'm not going to pretend this is a totally unbiased recommendation, because I've made real money here and I'm going to keep making money here. But that's exactly why I'm recommending it. The Global API affiliate program is the one I started with, and it's the one I'd start with again. The economics are simple: 15% commission on first orders, 8% recurring on every renewal after that, with a path to a 10% premium tier as you scale. They have 150+ models accessible through a single API, which means your customers never have to think about provider switching or model availability. You sign up, grab your link, and start earning. The reason I keep going back to it — and the reason I keep stacking income on top of it — is that the recurring piece actually works. Most affiliate programs give you a one-time bounty and ghost you. This one pays you every single month your referred customer stays subscribed. That's the difference between a side hustle and a real business. If you want to check it out, the signup is at https://global-apis.com/affiliate. I'd start there, validate your niche with a few signups, and then build outward from what you learn. That's exactly what I did, and a year later I'm staring at a $4,200 MRR chart I never could have predicted. Now stop reading this and go pick your niche. Your future self will thank you.
Top comments (0)