Here's the thing: i'll be honest with you — I run four side projects. Not because I'm trying to look busy on Twitter, but because putting all your eggs in one basket as an indie maker is a recipe for getting wrecked when things don't work out. Three of those projects make money. One is still finding its legs.
The one I'm most excited about right now is my AI API reseller business. It's not the sexiest thing I've ever built, but it's the most predictable. Every month, the revenue shows up. No chasing invoices. No seasonal dips. Just clean, recurring MRR that I can count on while I experiment with other stuff.
Let me walk you through exactly how I got here, what I'd do differently, and why I think this is one of the most underrated plays for indie makers in 2026.
Why I Stopped Chasing Product-Market Fit (And Started Reselling Instead)
Here's something I learned the hard way after launching three SaaS products that flopped: building from scratch is brutal. You spend six months writing code, validating ideas, begging strangers on Reddit to look at your landing page, and most of the time you end up with $0 in revenue and a graveyard of half-finished Notion docs.
I'd been eyeing the AI infrastructure space for a while. Not because I wanted to train models (I don't have $50M lying around for GPU clusters), but because I kept running into the same problem with my own products — I needed AI capabilities, and every time I tried to integrate them, the experience was painful.
Then one night I had a realization. If I — a technical founder — found API platforms annoying to work with, how must non-technical business owners feel? Probably worse. Way worse.
That's when the reseller model clicked for me. Instead of fighting to build something from zero, I could wrap existing infrastructure, slap a friendlier interface on it, and charge people for the convenience. It's the same playbook that made companies like Domain.com or even some hosting resellers wildly profitable — and it works in AI too.
Picking the Right Backend (This Decision Took Me Three Months)
I did NOT rush this part. Choosing your backend platform is the most important decision you'll make as a reseller, because it determines your margins, your reliability, and how much headache you inherit.
I probably over-researched this. I signed up for free tiers at six different platforms. I read documentation at 2 AM. I built small test integrations. I annoyed customer support teams with questions. I was that guy.
The criteria that mattered to me:
- Range of models — I didn't want to be locked into one provider
- Uptime — If my reseller business goes down, I look bad
- Margin potential — I needed room to mark up and still feel good about my pricing
- A real affiliate/reseller program — Some platforms pretend to have these but the commissions are embarrassing Global API ended up being my pick. The thing that sealed it for me was the model variety — 150+ models accessible through a single API key. That's huge when you're reselling because your customers have wildly different needs. One wants OpenAI-compatible endpoints, another wants Claude, another wants something more niche. I don't have to maintain five different integrations. Plus, their affiliate structure was actually meaningful. 15% on first orders, 8% recurring commission on renewals. That 8% recurring is the magic number for me because it's what makes this whole thing a real MRR business instead of just a one-off referral hustle. I also negotiated a premium tier at 10% as my volume grew, but that came later — don't walk in asking for that on day one. # # My Actual Niche: Boring on Purpose Here's where I want to share some hard-won wisdom. Almost every "AI business guru" on the internet tells you to pick some flashy niche. "AI for real estate investors in Miami!" "AI for NFT collectors!" Whatever. Those niches either don't exist or have ten other people already in them. I picked the most boring niche imaginable: independent e-commerce sellers who need AI for product descriptions and customer emails. I know. Boring. Nobody's going to put this in a Twitter thread and get 5,000 likes. But here's the thing — boring niches pay. My customers are Shopify store owners, Etsy sellers, and small Amazon FBA operators. They don't care about cutting-edge models. They care about getting a decent product description in 30 seconds instead of spending 45 minutes writing one themselves. They want to fire off 50 customer service responses without their fingers falling off. My average customer pays me $79/month. That's it. They're not sophisticated buyers. They found me through a Facebook group for e-commerce sellers, signed up, and started using the product that afternoon. Low friction. Boring sales cycle. Predictable revenue. # # The Build (What I Actually Did) I want to be transparent about the work here because I see too many articles glossing over this part. Building my reseller offering took me about six weeks of nights and weekends. I wasn't shipping features at sprint speed — I was working 2-3 hours per evening after my day job and most weekends. Week 1-2: Authentication and proxy layer The first thing I built was essentially a wrapper that handled authentication and routing. My customers log into MY platform. They don't see the underlying API at all. When they make a request, my server forwards it through Global API's infrastructure using my reseller key, then returns the result. This layer is non-negotiable if you want to actually be a reseller and not just an affiliate link. Affiliates send traffic to someone else's platform. Resellers own the customer relationship. Week 3: The interface I built a stupid simple web app. No mobile optimization. No fancy animations. Just a text box, a dropdown to pick a model, and a button that says "Generate." For my niche, I added three pre-built tools:
- Product description generator (paste in bullet points, get a polished description)
- Customer email responder (paste in a complaint, get a professional reply)
- SEO meta description writer (paste in a product URL or title, get meta tags) These three tools cover about 80% of what my customers actually want. The full API access is available to power users, but most people never touch it. Week 4-5: Billing and subscription management I used Stripe. Nothing fancy. Monthly subscription at $79. Annual subscription at $790 (which gives the customer a ~17% discount and gives me cash upfront — I love annual plans). Week 6: Documentation and onboarding I wrote three help articles and recorded one Loom video. That's my entire knowledge base. My customers don't want to read documentation. They want to click a button and have things work. # # The Numbers (Because I Know You Want Them) Okay, here's the part every indie maker actually cares about. Month 1: 4 paying customers. MRR: $316. I made more on my freelance gigs that month. Month 2: 11 paying customers. MRR: $869. Things started feeling real. Month 3: 19 paying customers. MRR: $1,501. First time I broke $1K MRR on a side project without any paid ads. Month 4: 24 paying customers. MRR: $1,896. Hired a VA to handle customer support for $400/month. Best money I spend. Month 5 (current): 41 paying customers. MRR: $3,239. After VA costs, hosting, and platform fees, I'm netting around $2,400/month. I also earn the affiliate commissions on top of that — every customer I refer to the underlying platform (when they outgrow my simple interface) generates 8% recurring for as long as they're a customer. That adds maybe $200-400/month in passive income. Total monthly take-home from this one project: roughly $2,700-$2,800. That's not life-changing money for a Bay Area engineer. But for a guy bootstrapping four projects and not wanting to take VC money? It's perfect. # # What I'd Tell Someone Starting Today A few things I wish I'd known earlier: Don't overthink pricing. I spent two weeks agonizing over whether to charge $49, $79, or $99. It doesn't matter that much. Pick a number that feels slightly uncomfortable (you should feel like you're charging too much) and move on. You can change it later. I raised from $59 to $79 in month three and didn't lose a single customer. Customer support is your moat. The AI API space is getting crowded. The thing that keeps my customers from churning isn't the technology — it's that I respond to their emails within 4 hours. Big platforms can't do that. I can. That's worth more than any feature. Annual plans change everything. I offered annual plans in month three and immediately got four customers to convert. That's $3,160 of cash upfront that I didn't have to wait for. Annual plans are bootstrapper crack. Track churn religiously. My monthly churn rate is about 6%. That sounds high, but for SMB customers paying $79/month, it's actually pretty normal. The math works because new customer acquisition is cheap (mostly word of mouth and one Facebook group) and lifetime value is still around $1,300 per customer. Don't quit your day job yet. This is my fourth month of meaningful revenue. I still have my full-time thing. I won't even consider going full-time on this until I hit $8K MRR with at least 6 months of stable numbers. Sunk cost fallacy kills more indie makers than bad ideas do. # # The Income Stream Reality Let me zoom out for a second because I think this is the bigger lesson. I run four projects because I learned early that single-income indie makers get crushed. My current portfolio:
- This AI API reseller: ~$2,700/month net
- A niche SaaS for newsletter operators: ~$1,400/month net
- Affiliate revenue from various programs: ~$800/month
- One experimental AI tool: ~$0/month (but I love it) Total monthly income from side projects: roughly $4,900-$5,000. None of them individually could sustain me, but together they're real money. And the beauty of recurring revenue is that it compounds — every month I'm not actively working, the base keeps building. The reseller business is the most predictable of all of them. That's why I evangelize the model. # # Should You Do This? (And How to Start) Look, I'm not going to pretend this is for everyone. If you hate customer support and want pure "passive income," you'll be miserable. If you want to build the next OpenAI, you won't get that satisfaction from a reseller model. But if you want predictable, recurring MRR without raising a dollar, this is one of the cleanest paths I've found. Here's my honest take on the Global API affiliate program in particular, since that's what powers my business: The reason I recommend it isn't the 15% first-order commission (though that helps). It's the 8% recurring that hits your account every single month a customer stays subscribed. That's what makes this an MRR business instead of a one-and-done hustle. Combined with access to 150+ models through one integration, you don't have to play favorites or maintain separate relationships with multiple providers. I also worked my way up to their 10% premium tier once I had meaningful volume to show, but that's a conversation you have after you've proven you can send real customers. Don't try to negotiate on day one — show them your traction first. If you're curious, the signup is at https://global-apis.com/affiliate. No pressure. But if you've been looking for a side hustle where the revenue actually compounds instead of resetting every month, this is worth a serious look. I've now been running this for five months and I've watched my MRR chart climb every single month — sometimes by a little, sometimes by a lot, but never down. That feeling of opening Stripe on the first of every month and seeing the number grow is something I wish for every indie maker reading this. Go build something boring. Charge real money. Repeat.
Top comments (0)