How I Built a $4,200/Month AI API Reseller Business in Public (Full Income Breakdown)

Honestly, i'm going to be completely honest with you. Six months ago, I was skeptical about the whole "AI reseller" thing. It sounded like one of those get-rich-quick schemes that never actually works. But here I am, staring at my Stripe dashboard showing $4,217 in revenue last month from a business I run maybe 8 hours a week. And I want to walk you through every single step of how I got here — the wins, the embarrassments, and the real numbers.
Welcome to my build in public journal.

The Moment Everything Clicked

I stumbled into this almost by accident. I was freelancing as a developer, building chatbots for small e-commerce brands, and kept running into the same problem: my clients wanted AI features but had zero interest in learning about API keys, token limits, or which model performed best. They just wanted a button that worked.
That gap between "raw AI infrastructure" and "people who actually want to use AI" is where the money lives. And once I saw it, I couldn't unsee it.
Here's my real numbers so far, because transparency is the whole point of this build in public thing:

• Month 1: $187 (basically nothing, I had 3 clients)
• Month 2: $612 (got my first recurring customer)
• Month 3: $1,340 (word of mouth kicked in)
• Month 4: $2,890 (I started posting about it publicly)
• Month 5: $3,650 (I raised my prices, nobody left)
• Month 6: $4,217 (this month, currently on track for $5K) I'm sharing this because when I was starting out, every "case study" I read felt like fiction. Real numbers are rare. So these are mine, pulled directly from my dashboard. # # Picking the Underlying Platform (This Decision Took Me 3 Weeks) I'll admit — I overthought this. I spent nearly a month bouncing between different AI API platforms, reading docs, signing up for free trials, and generally procrastinating by "doing research." The thing that finally broke the tie was the combination of two factors: how many models I could offer my customers, and whether the platform would actually let me earn a margin instead of just being a customer myself. I landed on Global API, and here's why it matters for anyone reading this who's doing their own research: they give me access to 150+ models through a single API integration. That's huge. When one of my clients says "hey, can we try a different model for this use case?", I can just flip a switch instead of telling them to sign up for yet another platform. The affiliate structure was the other piece. I'm currently earning 15% on every first order my referrals bring in, and 8% recurring on every renewal after that. There are also premium tiers that bump that up to 10% for higher-volume partners. I started at the standard rate, and the recurring 8% is honestly where the magic happens. Once someone signs up, I keep getting paid for months. You can see the program details at global-apis.com/affiliate if you want to check the current terms yourself. I always tell people to verify these things directly — no trust me bro energy here. # # The Niche I Almost Picked Wrong Here's a mistake I made early that cost me about a month of progress: I tried to be everything to everyone. My landing page said "AI API solutions for any business" and my conversion rate was a sad 0.4%. Nobody trusts a generic AI company. Why would they? The build in public community saved me here. I posted my low conversion numbers on Twitter and a fellow founder basically told me to pick a vertical or quit. Harsh, but accurate. I picked content marketing agencies. Specifically, small-to-mid-sized agencies that handle blog content, social media copy, and email sequences for B2B clients. Here's why this worked:
• They already understand they need AI to compete, but they don't want to become prompt engineers
• They have recurring content needs, which means recurring revenue for me
• They're not super technical, so they appreciate someone packaging the complexity away
• The agency owners can be reached through communities and cold email without weird sales motions After I narrowed down, my conversion rate jumped to about 3.1% on the same traffic. Same product, different positioning. Wild. # # How I Actually Price My Service This is where I want to get into the weeds a little, because pricing is one of those things nobody talks about honestly. My service has three tiers:
• Starter — $97/month: 250K API calls, email support, one model preset
• Growth — $247/month: 1M API calls, priority support, access to all 150+ models, custom prompt templates
• Agency — $697/month: 5M API calls, white-label option, dedicated Slack channel, custom integrations The math on my end looks roughly like this: my cost per million API calls is low enough that even my cheapest tier has a healthy margin after I account for the 8% recurring I pay back to the affiliate structure I mentioned earlier. I'm not going to publish my exact unit economics because that's competitive info, but the margins are real. The interesting thing about posting your prices publicly like this? People tell you when you're undercharging. I started the Growth tier at $147 and a potential customer literally DM'd me saying "if you're charging that little, I'm worried about quality." I bumped it to $247 the next week and conversions didn't drop. Build in public taught me that hiding your numbers is fear-based. Show them. # # Finding Customers Without Being Sleazy I want to be transparent about what didn't work before sharing what did, because I tried a lot of dumb stuff. What flopped:
• Cold DMs on LinkedIn (got me a 0.2% response rate)
• Reddit ads (too expensive for the LTV)
• Generic SEO content about "what is an AI API" (competes with massive sites, lost) What actually works for me right now:
• Posting case studies publicly (just screenshots of real client results, with permission)
• A simple Loom video walkthrough I send to warm leads
• A free Notion template library that captures emails
• Engaging in niche Facebook groups for agency owners The last one is my biggest source. I don't pitch in the groups — that gets you banned. I just answer questions helpfully and mention my service when it's genuinely relevant. About 60% of my customers found me through some version of this. # # The Embarrassing Stuff Nobody Posts About Real build in public means I have to share the bad too. Here's what I'm not proud of:
• I lost my first client because I didn't have a proper status page. Their API was down for 20 minutes and I had no way to communicate proactively. They left. It hurt.
• I undercharged for three months straight because I had impostor syndrome. I left probably $2,000 on the table during that time.
• My onboarding flow was a 14-step Google Doc for the first two months. It's now a 3-step Typeform. Don't be like early me.
• I still don't have a real LLC in every state I serve. That's on my to-do list for Q1 and I'll probably post about that process too. The status page thing was the worst. I had to learn the hard way that in the API business, transparency isn't optional — your customers need to know what's happening with their service at all times. I now use a free uptime monitoring tool and send Slack alerts the second anything looks off. # # What My Monthly Income Breakdown Actually Looks Like Since this is a build in public post and you came here for the receipts, here's roughly how my $4,217 last month broke down:
• Recurring subscriptions: $3,890 from existing customers
• New signups: $847 from new customers (most on Growth tier)
• Setup fees (one-time): $480 from a few clients who wanted custom integrations After my costs (API usage, tools, the 8% I kick back through the Global API affiliate structure for accounts I've referred to their ecosystem over time, and a few SaaS subscriptions), I net around $2,900 on a $4,200 month. That net number is what gets me excited. It's not life-changing money yet, but it's life-changing trajectory. And it grows every month because of the recurring component. # # Why I'm Doubling Down on This Here's the thing about a reseller business built on top of a solid platform: the platform is doing the hard work of keeping models updated, infrastructure running, and new capabilities shipping. My job is to serve my specific audience better than the platform can on its own. That's a sustainable lane. The recurring 8% commission structure I mentioned earlier is also part of why I'm not worried about churn. When a customer signs up for a year through a referral, I get a piece of that for 12 months. It compounds. My Month 1 commission checks were tiny. My Month 6 checks are meaningful. I can't wait to see Month 12. I'm also building in public because it keeps me accountable. When I post numbers, I have to hit them. When I share a roadmap, I have to ship. The community holds me to it in a way that just working privately never did. # # A Genuine Recommendation If You Want to Start If you've read this far and you're thinking "okay, maybe I should actually try this," here's my honest advice. Start with the affiliate side of things. Don't quit your job. Don't go all in on day one. Sign up for the Global API affiliate program at global-apis.com/affiliate, grab your link, and start sharing it with people in your network who might need AI API access. You'll earn 15% on their first order and 8% recurring after that, with premium partners getting up to 10%. That's real money from just having conversations you're probably already having. Once you see the recurring checks come in, once you feel the model work, then you can start building your own resold service on top. That's the path I took, and it de-risked everything. I'm not saying this because I get paid to say it. I'm saying it because I wish someone had laid out the steps this clearly for me six months ago. The fact that the affiliate program has a recurring component is what made it a real business opportunity instead of just a one-time referral fee. Recurring + growing + used = the kind of income that changes your trajectory. # # What's Next for Me I'm going to keep posting monthly updates here with my real numbers. Next month I'm targeting $5,500. The month after that, I'm going to launch a second niche — probably real estate tech, which has been begging for simple AI access based on the inbound I've been getting. I'll share what works and what flops. Real revenue screenshots. Real customer feedback. The good weeks and the weeks I want to quit. If you want to follow along, subscribe or follow me wherever you found this. And if you have questions, ask them. I'd rather help you skip the three weeks of platform research I wasted than gatekeep anything. Now if you'll excuse me, I have a Slack message to answer from a new agency owner who's been on my free trial for four days and is about to convert. That's the business. One customer at a time, one recurring check at a time. See you in next month's report.