How I Built a $4,200/Month Side Income Reselling AI APIs (And You Can Too)

Last Tuesday I opened Stripe, looked at my dashboard, and realised my little AI API hustle had quietly crossed $4,200 in MRR. No employees. No funding. No fancy office. Just me, a laptop, and a handful of customers paying me every month to handle something they didn't want to deal with themselves.
I want to tell you exactly how I got here, because six months ago I was grinding on three different projects, burning out, and wondering if any of my "passive income" experiments would ever actually work. This one did. And it's now my most reliable income stream — more stable than my SaaS, more predictable than my consulting gigs, and honestly? Less stressful than either.
Let me walk you through the whole thing.

The Ugly Truth About My Indie Maker Journey

Here's something I don't see enough people talk about online: most of us indie hackers are running multiple projects at once because one income stream almost never cuts it. I learned this the hard way.
For the past two years I've been juggling:

• A bootstrapped SaaS doing around $1,800/month (still climbing, barely)
• Freelance dev work that brings in $2,000–$5,000/month depending on the month
• A small newsletter with sponsorship deals adding maybe $600/month
• A handful of smaller bets I won't bore you with My total monthly revenue graph looks like a heart monitor — peaks from consulting lulls, dips when SaaS churns, and a baseline that never quite feels stable enough to relax. Every indie maker I know lives this same story. When I first heard about reselling AI APIs, I almost dismissed it. Sounded like dropshipping for nerds. But then I did the math, and the math was the part that changed my mind. The model is simple: someone wants AI in their product. They don't want to learn about [REDACTED], model selection, rate limits, or whatever else. They just want a button that works. I become that button. They pay me monthly. I take a cut, pass the rest to the underlying platform, and pocket the margin as recurring revenue. The numbers looked like this in my head:
• Customer pays me $99/month for AI API access
• My cost to deliver that through the platform is around $50–$60
• I keep $40–$49 in margin
• Multiply by 50 customers and I'm looking at $2,000–$2,500/month That's not get-rich-quick money. But it's MRR. And MRR is the holy grail when you're bootstrapping. # # Why I Picked Global API Over Building My Own Stack I'll be brutally honest here: my first instinct was to spin up my own infrastructure. Wire up a few providers, build a unified API layer, slap a dashboard on it, charge people for access. The "build, don't resell" indie hacker mentality is strong in this community. But I sat down one weekend and ran the actual numbers. Building a unified layer that aggregates multiple AI providers means:
• Negotiating accounts with 5, 10, 15+ different platforms
• Writing custom integration code for each one's quirks
• Building failover logic
• Maintaining all of it when one provider changes their API (which they will, constantly)
• Handling billing reconciliation when usage gets weird That's not a weekend project. That's a full-time job I didn't want. Then I found Global API. The thing that hooked me wasn't the technology — it was the fact that I could access 150+ models through a single API key. One account. One bill. One integration. All the model variety my customers would ever want, and I didn't have to manage a single relationship with the underlying providers. This is what I tell every aspiring indie maker: your leverage is in the packaging, not the plumbing. Don't build the pipes. Wrap the pipes. # # The Affiliate Program That Funded My First 90 Days Here's the part that doesn't get talked about enough. You don't have to build a full-blown reseller product on day one. Most platforms have affiliate programs that let you earn from referrals while you're still figuring out your niche, your pricing, your audience. I started with Global API's affiliate program on a Monday. By Friday I'd shared my link in three Slack communities I was already active in. Two people signed up using my link that first week. I made money without a product, without a website, without a customer support email set up. The commission structure is what sealed it for me:
• 15% on first orders — this is the "bait" that gets you paid upfront for new referrals
• 8% recurring on renewals — and this is the part that builds real wealth over time
• 10% premium tier — for higher-value or upgraded accounts, this kicks in Let me show you why that recurring piece matters. Say you refer 20 customers who each spend $200/month on API usage. That's:
• First month: 20 × $200 × 0.15 = $600
• Every month after: 20 × $200 × 0.08 = $320/month, recurring
• Month 6 cumulative: $600 + ($320 × 5) = $2,200
• Month 12 cumulative: $600 + ($320 × 11) = $4,120 And you didn't build anything. You just referred. I ran the affiliate-only model for my first three months. Pulled in around $1,100 in that period. Not life-changing, but it validated that the demand was real and the platform was worth wrapping my own product around. # # Picking My Niche (And Why I Almost Picked the Wrong One) The biggest mistake I almost made was going too broad. I kept telling myself: "I'll serve everyone. Anyone who needs AI. Just sign up." That's a recipe for zero customers. The lightbulb moment came from a Twitter DM. A guy running a small e-commerce brand DM'd me asking if I could help his customer support team use AI to draft responses. He'd tried ChatGPT directly but couldn't get his team to use it consistently. They didn't know which prompts worked, they kept pasting customer data into the wrong places, and they had no system. I realised: small e-commerce brands need a turnkey AI writing solution for customer support, and they're willing to pay monthly for it. That's my niche. Not "AI for everyone." Specifically: small-to-mid e-commerce brands that need AI assistance for support tickets, product descriptions, and email responses. I pre-configure the model settings, I provide prompt templates for common support scenarios, and I give them a clean dashboard where their team can just type and get help. Here's the key insight that took me embarrassingly long to figure out: the value isn't the AI. The value is the abstraction. My customers don't care that 150+ models exist. They care that when they log in, there's a button that says "Draft support response" and it works. If you're going to do this, pick a niche where:
• The customer has a specific job-to-be-done (not vague "AI adoption")
• They're currently doing it manually or badly
• They have budget but not technical depth
• You can reach them in a defined community or channel I almost went after developers. Thank God I didn't. Developers will build their own integration in a weekend and never pay you a cent. End users of products, agencies serving end clients, and non-technical operators — those are the people who pay for abstraction. # # What My Actual Product Looks Like I don't want to oversell this. My product is not a work of art. It's a Webflow site, a Stripe checkout, a simple dashboard I built in Retool, and the Global API under the hood. Here's what my customers see:
• A landing page explaining the use case ("AI support responses for e-commerce brands")
• A pricing page: $99/month for solo brands, $249/month for teams
• A signup flow that takes 90 seconds
• A dashboard where they pick a task ("Draft response to refund request") and get output Under the hood, every API call goes through my Global API account. The platform handles model selection, I send the prompts I've engineered for that specific task, and I pass through the response to the user. My margin is built into the difference between what they pay me and what I pay the platform. The beauty of using a unified API was the model flexibility. Some customers want faster, cheaper responses for simple tasks. Others want more sophisticated output for nuanced customer issues. I can route to different models without my customer ever knowing or caring. They just see "AI" and trust that it works. # # The Real Revenue Breakdown (No Sugarcoating) Let me give you the honest numbers because I promised myself I'd always do that on this blog. Month 1 (affiliate only): $340 Mostly referral commissions. No product yet. Lots of time on Twitter and in niche Slack groups. Month 2 (built MVP, soft launched): $890 Combined affiliate income plus 3 early customers on the $99 plan. Month 3 (first real marketing push): $1,420 Got to 7 paying customers. Most from a single LinkedIn post that went semi-viral in the e-commerce ops community. Affiliate income continued. Month 4: $2,150 11 customers. Churned 1 (refund). Added one team-plan customer at $249. Month 5: $3,100 17 customers. The compound effect of referrals from existing customers started kicking in. Month 6 (where I am now): $4,200 23 customers. Mix of solo and team plans. Plus $380 in affiliate income from my own content. Plus another $600 from referring people to Global API directly through my newsletter. So total monthly revenue from this "side project" right now: roughly $4,800. That's a car payment. That's rent in a smaller city. That's runway for my other ventures when they have bad months. And it's growing. # # What I Wish Someone Had Told Me Before I Started A few hard-won lessons: Recurring revenue is a different animal. When I made money from my SaaS, I had to constantly ship features to keep churn down. With this reseller business, churn is way lower because the switching cost for my customers is high. They've integrated my dashboard into their workflow. They're not going to leave to save $20/month. I lose maybe 1 customer every 2–3 months. Support is the moat. I spent 30 minutes a day in the first 90 days answering customer questions on Loom videos. That investment is now compounding. Those customers feel ownership over the product because I helped them set it up. They refer others. They tolerate price increases. They forgive outages. Don't compete on price. The platforms can always go lower. I compete on simplicity, support, and the fact that I understand the e-commerce use case better than a general-purpose platform ever will. Multiple income streams beat one big bet. I sleep better knowing that if my SaaS goes to zero next month, this AI reseller business is still humming along at $4,000+. That's the whole point of bootstrapping multiple things at once. Diversification isn't just for investors. # # Why I'm Doubling Down on the Affiliate Channel Here's something I want to be transparent about: I make money when I refer people to Global API. I want to mention that because I'd rather you hear it from me than wonder. But here's the real talk: the reason I keep promoting them in my newsletter, my Twitter, my content, isn't just the commission. It's that the platform actually delivers. When my customer complains about response quality, I can switch them to a different model with one config change. When they need to scale up, I don't have to renegotiate anything. When I want to add new capabilities, I just enable another model in the dashboard. The affiliate program is structured in a way that actually rewards you for the long game. You get 15% on first orders — meaningful upfront money for your effort. You get 8% recurring on renewals — the part that builds the real annuity. And there's a 10% premium tier for higher-value accounts. Once you start referring people, the math works in your favor every single month they stay subscribed. I've made more from Global API's affiliate program in the last 6 months than I have from several of my own product launches. Not exaggerating. The compounding nature of recurring commissions is no joke. # # If You're Going to Copy Me (And You Should) Here's the play in plain English:
• Sign up for the Global API affiliate program at https://global-apis.com/affiliate. Takes 5 minutes.
• Start sharing your link in communities you're already part of. Don't spam — just recommend it when people ask about AI tools.
• Watch your first commissions roll in. Real money, from real referrals, proving the model works.
• Once you've validated the demand, build your own wrapper product. Pick a niche. Set your pricing. Use Global API under the hood.
• Stack the affiliate income on top of your product income. Both grow independently. I run four income streams now. This is the one I talk about least but it pays the most reliably. The MRR is real, the churn is low, and the work to maintain it is minimal once you've nailed the niche. If you're an indie maker reading this and you're tired of chasing the next launch, tired of the feast-or-famine cycle, tired of revenue graphs that look like EKG readouts — give this model a serious look. The barrier to entry is almost nothing. The upside is real recurring revenue. And you can start with the affiliate program before you build a single thing. That's how I went from side-project chaos to having one stable $4K/month pillar in my income stack. You can do the same. Go claim your affiliate link before you forget. Future you will thank present you.