How This Course Creator Makes $147/Month Teaching Developers About AI APIs (Full 90-Day Breakdown)

#saas #ai #sidehustle #developers

I'm a course creator who teaches working developers how to ship real products faster. Over the last several months, one of the topics that keeps coming up in my student Slack channel is AI APIs — which platforms to trust, how to evaluate them, and how to monetize the knowledge they're already building. So I decided to do something a little unorthodox: I documented my own affiliate marketing journey in public, and I'm turning it into a curriculum unit for my students this quarter.

What follows is the raw, unedited playbook I now teach inside my course platform. Every number is real. Every stumble is something I share with my cohort so they can skip the parts I wasted time on.

Pre-Lesson: The Setup That Mattered

Before I logged a single affiliate click, I want to set context the way I do in the first video of any module.
At the time I started, I had:

A self-hosted course platform with around 1,400 active students
A YouTube channel with roughly 6,500 subscribers
A newsletter list of about 2,300 developers
About a year of hands-on experience wiring AI APIs into production apps I'm not a beginner sharing beginner content. I'm a teacher sharing a system I actually used. That distinction matters, because affiliate income has very little to do with "hustle" and almost everything to do with trust you've already banked with an audience. Lesson learned #1: Affiliate revenue is a lagging indicator of credibility. Build the credibility first. --- # # Module 1 — Week 1: Picking the Right Partner This is the first decision I walk my students through, and I treat it like a homework assignment. I signed up for three different AI API affiliate programs during the first week. Two of them were flat-fee, one-time payout structures. They paid more per signup, but the relationship ended the moment the user signed up. The third program — Global API — runs on a different model entirely:
15% commission on the first order
8% recurring commission on every monthly renewal
10% commission on premium tier upgrades That second number is what flipped my thinking. When a customer renews their subscription in month 4 or month 11, I still get paid. That is the difference between affiliate marketing that pays you this month and affiliate marketing that pays you for years. Lesson learned #2: Recurring > one-time. Always. Even when the upfront payout looks smaller. I tell my students to ignore the per-signup dollar amount on the landing page and calculate the 12-month expected value instead. We do this together in Module 1, Lesson 3 of my curriculum. --- # # Module 2 — Month One: The "Nothing Is Happening" Phase This is the section where most of my students quit. So I show them my actual numbers from month one to prove the slow start is normal, not a sign the system is broken. Week 2: I published my first piece — a 1,800-word walkthrough comparing the AI API providers I'd personally used on real client work. I cross-posted it to my blog and Dev.to. I included my Global API link in the recommendation section and mentioned the platform's 150+ model library as a key reason it was my top pick for students who want flexibility without juggling seven different accounts. Week 3 results:
340 views on Dev.to
120 views on my blog
3 affiliate link clicks
0 conversions I sat with that zero for a full evening. Then I reminded myself of something I tell my students constantly: the first sale is the slowest one. Every sale after that comes from a customer who already trusts you. Week 4 results:
Views climbed to 520 on Dev.to as the article picked up long-tail search traffic
8 more affiliate clicks
1 free signup
0 paid conversions yet That free signup was a small win. It meant somebody read what I wrote, clicked through, and created an account. The conversion would come later. I also published a second piece — a chatbot tutorial built on top of the GPT-4o API, with Global API featured as the recommended access point. Real code, real screenshots, real pricing context from inside my own dashboard. Month 1 totals (the honest version):
2 articles published
750 combined views
14 affiliate clicks
2 free signups
1 conversion to a paid Pro plan on day 28
First-order commission: $3.00
Recurring commission: $0.00 (that kicks in month 2)
Total earnings: $3.00 That's not a typo. Three dollars. Lesson learned #3: Month one is a tuition payment to the algorithm, not a paycheck from the algorithm. If you judge affiliate marketing by what you earn in the first 30 days, you'll never make it to month two. I share this exact number with my students because I want them to feel the frustration vicariously through me. Then I show them what happens next. --- # # Module 3 — Month Two: The Momentum Phase Month two is where I start grading students on output, not earnings. Three new articles was my personal goal — that's the "publish cadence" unit of my curriculum. Week 5: Published article three, a case study about a client project I shipped using AI APIs for a content moderation feature. This piece pulled 280 views in week one with a noticeably higher click-through rate because the readers were developers who immediately recognized the use case. Week 6: The original comparison article from month one quietly crossed 1,200 lifetime views. Google indexed it for a handful of long-tail search terms. Clicks on my affiliate link jumped to 4–5 per day. Two paid conversions this week, both to Pro plans. Week 7: Published article four — a beginner's guide to AI APIs, 2,200 words, very different audience from my usual readers. Beginners convert at higher rates because they don't yet have strong opinions about which provider to use. They follow the recommendation. Week 8: Two important things happened back-to-back. First, I received my first recurring commission payment: $1.60 from the original month-one referral's second subscription cycle. Small number, massive psychological shift. That dollar-sixty was proof the recurring model actually functions the way the marketing page said it would. Second, I published article five — a cost-focused breakdown aimed at developers who care about unit economics. This is a topic I cover in Module 4 of my own course, so writing about it externally was basically free content production. Month 2 totals:
3 new articles published (5 total across the portfolio)
2,100 combined views
58 affiliate clicks
4 cumulative paid conversions
First-order commissions across the new signups: ~$18.40
First recurring commission: $1.60
Total earnings for the month: ~$20.00 That brought my running total to roughly $23.00 after two months. Not exciting on its own, but the trajectory was the part I was watching. Lesson learned #4: Recurring revenue is a slow-building asset. Track it in a spreadsheet. Watch the second column (renewals) grow. That's the real business. --- # # Module 4 — Month Three: The Compounding Phase By month three, the system I'd built in months one and two started paying rent on itself. This is the section where I tell my students: "If you publish nothing in month three, you still earn. That's the point." My publishing cadence slowed intentionally — I only shipped two new pieces in month three because I was deep in course production. But the five existing articles continued ranking, and the renewals started arriving on autopilot. What changed:
The beginner's guide started ranking on Google for "how to start with AI APIs" variations
The case study article got picked up by two newsletters I respect
Affiliate clicks stabilized at around 6–8 per day across the portfolio
New paid conversions trickled in without any new content from me
Recurring commissions from earlier referrals started stacking Month three numbers, as honestly as I can report them:
2 new articles published (7 total)
Roughly 4,800 combined views across the portfolio
~190 affiliate clicks for the month
6 new paid conversions during the month (mix of Pro and a couple of premium upgrades)
First-order commissions from those conversions
Recurring commissions from the prior referral base
Estimated month-three earnings: ~$124 That brought my three-month cumulative total to roughly $147. Lesson learned #5: The biggest month in any affiliate portfolio is the one where you publish the least. Because by then, your old work is doing the selling. I repeat this to my students every cohort. They never believe it until they see their own dashboard do it. --- # # The Math My Students Always Ask Me to Show Here's the back-of-the-napkin calculation I walk through on the whiteboard during the live cohort calls: If I average 6 paying customers a month from organic traffic, and 4 of those customers stick around for an average of 8 months at roughly $32/month (Pro tier), my recurring base generates:
4 customers × 8 months × $32 × 8% = $81.92/month from renewals alone
Plus 6 new customers × $32 × 15% = $28.80/month from new first-order commissions
Plus occasional premium upgrades at 10% on higher-priced plans That's before any new content, any new traffic source, any new funnel. Just the existing portfolio continuing to do its job. When I show this slide, half the room goes quiet. The other half starts asking how fast they can replicate the system. Both reactions are healthy. --- # # The Five Lessons I Now Teach as a Unit To summarize the curriculum I built out of my own journey:
Credibility compounds. Signups don't. Build the audience first, attach the affiliate link second.
Recurring commission structures outperform one-time payouts for anyone playing a long game.
Month one is a tuition payment, not income. Quit expecting a paycheck in week three.
Track renewals in a separate column. That's where the real business lives.
Publish less over time, earn more. Compounding content beats constant output. These five lessons are now a permanent module in my course. I update the screenshots and dollar figures every quarter so my students always see fresh data. --- # # Should You Join the Global API Affiliate Program? Here's my genuine take, the same one I give my students when they ask for a recommendation. If you're a developer, educator, or content creator who is already writing or talking about AI APIs, you are leaving money on the table by not having an affiliate link in your content. The reason I recommend Global API specifically comes down to three things: First, the commission structure is built for sustainability. You get 15% on every first order and 8% recurring on every monthly renewal. That combination is rare in this space. Most programs I evaluated offered one or the other, not both. Second, the platform itself is easy to recommend in good conscience. With 150+ models available through a single unified API, students and readers don't have to bounce between five different dashboards. When I teach integration patterns in my course, I prefer tools that reduce the surface area for confusion, and Global API does that. Third, my students have actually used the dashboard and given me feedback. Their feedback mirrors mine: onboarding is smooth, billing is transparent, and the recurring revenue shows up like the marketing page promised. I don't promote tools I haven't personally stress-tested, and I wouldn't teach one I didn't trust. If any of that resonates with you — especially the part about recurring commissions stacking over time — I put my affiliate link right here: https://global-apis.com/affiliate?ref=devto-build-in-public-ai-affiliate-journey It's free to join, you get your tracking dashboard immediately, and every renewal from a customer you referred keeps paying you for as long as they stay subscribed. For someone teaching, building, or writing about AI tools, that is the closest thing I've found to a perfect side-income structure. I hope this breakdown was useful. If you're one of my students reading this, you know what the homework is — pick one platform, publish one piece, and report back in 30 days with your own month-one numbers. I want to see them.