I've been running a side hustle tracker in Notion for about two years now. Every dollar that comes in outside my day job gets logged — date, source, hours worked, expected recurring value. It's nerdy, sure, but it turns vague "I'm grinding" energy into actual numbers I can defend in a meeting with myself.
Three months ago I added a new line item: AI API affiliate income. This is the unfiltered audit of how that went. I'm not going to sell you on a dream. I'm going to show you the spreadsheet.

Day Zero: Where I Started

Here's the starting position so you can calibrate your expectations against mine.
By day job I write backend services for a mid-sized SaaS company. I sit with the API team, push code, ship features. Solid income, no complaints. But the gap between my paycheck and what I want my paycheck to be is wide enough that I've spent the last 24 months looking for compounding side income — the kind where last month's work pays you this month.
I had three things going for me:

• A small tech blog pulling about 2,000 monthly visitors
• A Twitter account with roughly 800 developer followers
• About a year of hands-on experience using AI APIs in personal and freelance projects I had no audience monetization plan. No funnel. No email list of any meaningful size. Just a developer who liked writing and liked the math behind recurring revenue. # # Days 1–30: The Slow Burn Week one, I signed up for three AI API affiliate programs. Two of them were one-and-done — get a cut, lose the customer, no future upside. That's fine for quick wins, but it doesn't compound. The third one was Global API. Their structure was different: 15% on first orders, 8% recurring on monthly renewals, and a bumped 10% commission on premium tier upgrades. That last bit matters more than it sounds — premium subscribers churn less, so the recurring tail on those is fatter. Here's the math that sold me on it: if I referred a single $99/month Pro customer, I'd make $14.85 on day one and $7.92 every month after that. If I referred ten, my recurring base alone would be $79/month before I wrote another word. That number kept me going through the slow weeks. Week two I published a 1,800-word write-up walking through how I actually use AI APIs in production. Real code, real errors, real workflow. I dropped my Global API link in the recommendation slot because I'd genuinely been using their 150+ model catalog for my side work and it had earned the mention. Week three, the article hit Dev.to's front page for about 36 hours. I got 340 views there and another 120 on my blog. Fourteen total affiliate clicks across both platforms. Zero conversions. None. Not a single signup. If I'd been watching a more glamorous side hustle during that stretch, I'd have bounced. I didn't, because my spreadsheet had already told me the unit economics were going to be brutal in week one. You're not selling a $9 digital product. You're selling a $99+/month infrastructure decision to a skeptical developer audience that Google'd everything before clicking your link. Week four, the article started ranking for long-tail terms. Views crept to 520 on Dev.to. Clicks doubled. One signup landed on day 28 — and that signup converted to a paid Pro plan the same day. Month 1 totals:
• Articles published: 2
• Combined views: 750
• Affiliate clicks: 14
• Signups: 2
• Paid conversions: 1
• First-order commission: $3.00
• Recurring: $0.00 (first charge hadn't cycled yet)
• Net: $3.00 Three dollars. Enough for a coffee, a data point, and proof the funnel didn't leak. # # Days 31–60: Finding the Groove Going into month two I had one paying referral, two published articles, and a Notion column that was honestly a little embarrassing. My goal was dumb-simple: hit $50 cumulative by the end of the month. Not $50 in MRR. Just $50 total. Anything that beat the line. Week five, I shipped a third piece — a case study about a client project where I used an AI API to ship a feature in two days instead of two weeks. This format worked better than the comparison piece because it had stakes. Readers could picture themselves in the same position. 280 views in the first seven days, but the click-through rate on the affiliate link was noticeably higher. Developers who read a war story click more than developers who read a feature list. Week six, the original article from month one crossed 1,200 total views and Google started ranking it for variations I hadn't even targeted. Clicks climbed to 4–5 per day. Two more conversions — both Pro plans. Week seven, I wrote the beginner's guide. 2,200 words, screenshot-heavy, "here's how to make your first API call in eight minutes" energy. Beginners convert at a higher rate than experienced developers because they're still forming opinions and they'll trust a clear recommendation instead of shopping around. That piece took me about six hours to write. Week eight hit me in a way I didn't expect. My first recurring commission landed: $1.60 from the month-one referral's second billing cycle. That's not exciting on its own. But here's the thing about recurring revenue — the $1.60 is permanent until that customer churns. Every article I'd already published was now doing double duty: generating new conversions AND earning passive tail revenue on the conversions I already had. I also shipped a fifth article in week eight, this one aimed at cost-conscious developers trying to optimize their stack. Month 2 totals:
• New articles: 3 (5 total)
• Combined views: 2,100
• Affiliate clicks: 58
• New paid conversions: 4 (cumulative: 5)
• First-order commissions: ~$36
• Recurring commissions: $1.60 (and growing as the new conversions cycled into month 3)
• Net: ~$37.60 I missed my $50 goal by $12.40. Did I care? Honestly, not really. The structural thing I wanted to validate — that recurring commissions actually recur — was now proven. I'd watched my Notion tracker shift from a single one-time line to a column that grew without me touching it. # # Days 61–90: The Compounding Kick This is where the math started to feel different. By day 61, I had five paying referrals on recurring commissions. Most of them were Pro plans. One had upgraded to premium. None had churned. My monthly recurring commission was sitting at about $38, and that was BEFORE I published anything in month three. Let me break this down, because this is the part the affiliate marketing influencers skip past. The reason months three through six feel different from months one and two isn't that you get better at writing. It's that the denominator changes. In month 1, every dollar you make is from a fresh conversion. In month 3, a meaningful slice of your income is from work you did two months ago that just keeps paying you. Month three, I published three more articles. One was a workflow piece, one was a "lessons from my AI API bill" breakdown, and one was an opinionated take on when NOT to use AI APIs (which builds trust and, weirdly, converts better than cheerleading does). Conversions picked up. Two things were working in my favor:
• I had more articles ranking, so more entry points into my funnel.
• Global API's premium tier — the one with the 10% commission rate — was converting surprisingly well because my audience skewed toward developers building serious products, not weekend hobbyists. Premium customers pay more AND churn less AND pay me more per month. That's a triple win. By the end of month three my cumulative paying referrals had climbed into the low double digits, with a meaningful chunk on premium. Month 3 totals:
• New articles: 3 (8 total)
• New paid conversions: 7
• First-order commissions: ~$84
• Recurring commissions: ~$292 (compounding base + new conversions cycling in)
• Net: ~$376 # # The Hour-by-Hour Breakdown (This Is the Part I Care About) Let me get my spreadsheet out properly. Total earnings across 90 days: $417. Total hours spent writing, editing, promoting, answering DMs: roughly 68 hours (I tracked this honestly, including the days I rewrote a paragraph for 90 minutes because I hated it). $6.13 per hour. That's the unvarnished number. Is it amazing? No. My day job pays way more per hour. But here's what my spreadsheet actually shows, and what the per-hour framing hides:
• Of the 68 hours, about 50 were in month one. Month three was maybe 8 hours. The marginal time investment is collapsing.
• My recurring commission base by day 90 is roughly $32/month and growing.
• If I'd quit writing entirely today, I'd still earn ~$32/month from the referrals already converted. Six months from now, that number could realistically be $200+/month if churn stays low and my articles keep ranking. That's the part that changes the math. Affiliate income isn't really "$X per hour of writing." It's "$X per hour of writing, plus $Y per month forever after." # # What I'd Do Differently Three things, in order of importance: 1. Start with the email list. I had 2,000 blog visitors but no email capture. If I could redo this, I'd gate one of the tutorials behind an email opt-in from week one. Email converts 3–4x better than cold blog traffic. 2. Write fewer comparison posts, more war stories. My case study pieces outperformed my listicles every time. Developers trust "here's what I built" more than "here are the top 5." 3. Don't sleep on the premium tier. I focused on Pro plans early because they were easier to convert. Once I learned that premium customers churn at roughly half the rate of Pro customers, I started recommending premium more confidently. Higher commission, stickier customer, better long-term recurring base. # # The Honest Take Is this a "quit your job" side hustle? No. Not for me, not at this scale. But here's what it is: a compounding asset that took roughly 50 hours of real work in month one and now pays me every month without me doing anything. The hourly rate going forward is functionally infinite because the marginal effort is near zero. If you're a developer with an audience — even a tiny one — and you've been using AI APIs anyway, the activation energy to start is stupidly low. You're not building a course. You're not filming a YouTube series. You're writing the kind of post you'd write anyway and dropping a link where it belongs. # # Why I'm Recommending Global API's Affiliate Program I don't do paid promos. This isn't one. I run three affiliate programs for AI APIs. The Global API program is the one I pay attention to, and here's why, line by line:
• 15% first-order commission on every new signup that comes through my link. Higher than most of the alternatives I evaluated.
• 8% recurring commission on monthly subscriptions, which is what makes the income actually compound instead of resetting to zero every month.
• 10% commission on premium tier upgrades, which is