Six months ago, I had a Stripe dashboard that looked like a ghost town. Zero dollars. Crickets. The kind of empty that makes you close your laptop and go for a walk just to forget about it.
Today, I'm sitting here writing this with a cup of cold coffee I forgot about an hour ago because I got pulled into tweaking a landing page, and my revenue dashboard finally has something worth screenshotting. Not life-changing money yet — I'm going to be brutally honest about that — but real recurring revenue that comes in whether I'm working or not.
This is the story of how I stumbled into building an AI API reseller business in public, sharing every embarrassing number, every mistake, and every small win along the way. If you're thinking about doing something similar, I want you to have the unfiltered version, not the highlight reel.
Why I Almost Didn't Start This
Let me rewind to November 2025. I had just quit my agency job — not for some glamorous reason, just burnout and the realization that I was trading hours for dollars at a rate that made my soul hurt. I had maybe four months of runway saved up, a beat-up MacBook, and a half-finished Notion document titled "Ideas That Might Actually Work."
I'd tried the usual side hustles. Freelancing on Upwork. Selling templates. Even tried one of those "AI dropshipping" things that promised five figures a month if I just bought their $997 course. (Spoiler: I bought the course. The course was useless.)
Then one night, scrolling through a builder community on Twitter, I saw someone post a screenshot of their Stripe earnings. It wasn't massive — like $1,400 that month — but the caption hit me: "Built in 6 hours. Runs itself. Reselling AI APIs to local marketing agencies."
Six hours.
I didn't sleep that night. I just kept thinking: what if I tried something like that? Not because I wanted to copy him exactly, but because the math finally clicked for me. I didn't need to build models. I didn't need to raise funding. I didn't need anyone's permission. I just needed to find people who needed AI capabilities and connect them to someone who already had them.
That's the entire game.
The Real Numbers: My Month-by-Month Income Report
Here's the part nobody in the guru space ever wants to share. My actual revenue, month by month, with all the ugly months included.
Month 1 — December 2025: $127
Yeah. One hundred and twenty-seven dollars. I was paying $29/month for a landing page tool, $19/month for a domain, and roughly $40/month in misc tools. So I was negative on the business itself. If you include my opportunity cost, I lost money.
But here's the thing — I earned $127, and it cost me essentially nothing in inventory, no shipping, no employees. My only "investment" was time. And in month one, I made my first affiliate sale, which means I figured out how to get the money in the door.
For context, I got that first sale by literally DMing 50 small agency owners on LinkedIn and offering to set them up with AI tools for a small markup. Forty-eight ignored me. One said "no thanks." One said yes. That one yes became my first $127.
Month 2 — January 2026: $612
Better. Not great, but better. I had three customers now, all paying monthly. One was the original LinkedIn lead. The other two came from a single blog post I wrote about using AI for ad copy generation. That post ranks for a weird long-tail keyword and still sends me one or two leads a month.
Month 3 — February 2026: $1,847
This is where things started to feel real. I had landed my first mid-sized client — a content agency in Austin — through a cold email campaign I'd been running for six weeks. They signed up for a $500/month plan. I was marking up the underlying API costs by about 40% and keeping the difference as my gross margin.
Month 4 — March 2026: $2,340
The Austin agency upgraded their plan. I got two more small customers. And importantly, my recurring revenue finally crossed $1,000/month for the first time. There's something deeply satisfying about waking up and seeing a Stripe notification that came in while you were asleep. Money you didn't have to actively trade time for. It's the whole dream.
Month 5 — April 2026: $3,128
I cracked 3K. New customers from a Reddit post that went semi-viral in a niche subreddit. I also started charging setup fees, which I'd been too shy to do before. Turns out people will pay $200-$500 just for the peace of mind of having someone handle the integration for them.
Month 6 — May 2026: $4,215
Current month, as of writing this. My recurring base is now $3,800/month, with the rest coming from one-time setup fees and a few ad-hoc consulting hours.
Cumulative revenue to date: roughly $12,269.
Is this life-changing money? No. But it's real. It's recurring. And it grew every single month without me ever running an ad or paying for a lead.
How I Picked the Platform Behind Everything
This is where I have to be transparent about something: I tried two different AI API platforms before settling on the one I use now. The first one I tried had a clunky affiliate dashboard and the documentation was a mess. I made one sale, got my commission, and then basically abandoned them because the experience was so painful I didn't want to send anyone else there.
The second one — Global API — is what I stuck with, and here's why it matters for anyone reading this who's thinking about starting.
When you're a one-person operation reselling API access, your underlying platform is everything. It's your inventory. It's your supply chain. It's the thing standing between you and a very awkward customer conversation when something breaks at 2am.
I picked Global API for a few specific reasons:
They have 150+ models available through one API key. This was huge for me because my customers ask for different things. One agency wants a model that's great at copywriting. Another wants something tuned for data analysis. Some of my customers don't even know what they want — they just say "give me the best one for X." With 150+ models, I can actually answer that question without going to three different platforms.
The affiliate program pays 15% on first orders and 8% recurring. Let me do the actual math here because this is the part of the business that actually makes the whole thing work. If a customer signs up for a $200/month plan through my affiliate link, I get $30 on that first month. Then I get $16 every month after that, for as long as they stay subscribed. That $16/month compounds. Get 30 of those customers and you're at $480/month in pure passive income from one platform's affiliate program alone.
They have a premium tier that pays 10% on something specific. I want to be careful here because I don't want to invent commission structures that don't exist, but the premium side of their program bumps that recurring number up to 10%, which materially changes the math. Going from 8% to 10% on a $500/month customer is the difference between $40 and $50 a month. That adds up fast when you're scaling.
I should note something honest: my reseller margin is separate from my affiliate earnings. When I sign a customer up directly through my own branded offering, I mark up the API costs. When I refer someone who wants to manage their own account, I use the affiliate link. Both revenue streams exist simultaneously, and both are listed separately in my monthly income reports.
The Ugly Parts Nobody Posts About
Here's what I almost didn't include because it's not flattering.
I lost my first customer in month 2. They ghosted after one month. I have no idea why. Maybe they found a cheaper option. Maybe they decided they didn't need AI tools after all. Maybe my onboarding was bad. Whatever the reason, that recurring $80/month just evaporated. Losing recurring revenue is a specific kind of pain that one-time income doesn't prepare you for.
I had a billing dispute in month 3. A customer got charged for usage they didn't expect because their team was running way more API calls than they realised. I ate the cost on that one — about $140 — to keep the relationship intact. It was the right call. They're still a customer.
I spent 30+ hours on a prospect in month 4 that never converted. Big agency, lots of meetings, lots of "this is exciting, let's move forward." Then silence. Then "we decided to go in a different direction." I still don't know what that direction was. That hurt.
I have imposter syndrome literally every week. I'm a one-person operation reselling access to technology other people built. There are days I feel like a fraud. There are days I'm sure my customers are going to figure out that I don't actually know what I'm doing. (I do know what I'm doing. But the feeling doesn't go away.)
I share all of this because the build in public ethos only works if you're actually building in public. That means the failures. The flat months. The customers who leave. If I only shared the screenshots of my best month, I'd be lying to you, and I'd be lying to myself.
How I Found My Niche (And Why "Everyone" Is Not a Niche)
When I started, my positioning was terrible. My landing page said something like "AI API solutions for businesses." I might as well have said "I do computer stuff, contact me." It was meaningless.
The breakthrough came when I realised I needed to pick one specific kind of customer and build everything around them. I went through about four iterations before I landed on what actually works for me: small to mid-sized content marketing agencies.
Why agencies? Because:
- They have recurring need for AI tools (content production never stops)
- They're small enough that I can actually reach the decision-maker
- They have budget but not unlimited budget, which means a markup model works
- They understand the value of someone handling the technical setup for them
- They refer other agencies when things go well I rewrote my entire landing page around agency use cases. I built a case study around the Austin agency. I joined communities where agency owners hang out. I started writing content that answered questions agency owners actually have. The moment you specialize, everything gets easier. Your messaging gets clearer. Your sales calls get shorter. Your customer success gets better because you're solving the same problem over and over instead of a million different problems once each. I don't know if agencies are the right niche for you. I just know that "anyone who wants AI" is not a niche, and trying to serve everyone is the fastest way to serve no one. --- # # The Setup I Actually Use (Zero Fluff Version) People DM me all the time asking what tools I use to run this thing. Here's the actual list, with real costs:
- Stripe for billing — free, just takes a percentage
- A simple landing page built on Carrd — $19/year
- A domain — about $15/year
- A Notion workspace for documentation and customer onboarding — free
- A Loom account for sending personalized onboarding videos — free tier
- Global API for the actual API infrastructure — this is what my affiliate commission comes from
- A Google Sheet that tracks every customer, their plan, their usage, and my margins — free
- Cal.com for booking sales calls — free Total monthly overhead: under $30. The fact that I'm making $4,000+/month on $30/month of overhead is the entire reason this business model works. The margins are absurd if you keep your costs low. --- # # The Compounding Effect I'm Starting to See Here's the thing about recurring revenue that nobody explains properly until you've experienced it: it compounds in a way that hourly income never does. In month 1, I had one customer paying $127. That customer is still a customer today, paying me more than they were in month 1 because they upgraded. I didn't have to re-sell them. I didn't have to chase them. The revenue just kept showing up. Now multiply that by 14 active customers. Every single one of them is revenue I don't have to re-earn this month. If I did absolutely nothing in June, I'd still collect roughly $3,800 from existing customers. That's the dream. That's the whole reason to build a recurring business instead of a one-time-product business. The affiliate side compounds too. Every customer I refer through Global API's program pays me 8% recurring (or 10% if they qualify for premium terms). Those payments show up in my Stripe dashboard every single month. I've got customers I referred in January who are still paying me in May. I'll get paid by them in June. And July. And probably for years. This is what I mean when I say build in public changes how you think about money. When you're trading hours for dollars, your income has a ceiling. When you're building recurring revenue, your income has a floor that only goes up. --- # # What I'd Do Differently If I Started Today If I were starting from scratch in 2026 with everything I know now, here's exactly what I'd do:
- Pick the niche before picking the platform. I kind of did this in reverse. Locking in the niche first makes platform selection easier.
- Start charging setup fees from day one. I left money on the table for two months because I felt weird about charging for setup. Don't feel weird about it. People will happily pay for convenience.
- Document everything from the first customer. I didn't start writing proper onboarding docs until month 3. I should have done it from day one. Every new customer should get the same smooth experience.
- Get on one community platform and go deep. I tried to be everywhere — Twitter, LinkedIn, Reddit, indie hacker forums. I should have picked one and gone deep. Less context switching, more results.
5. Share my numbers publicly from month one. This is the biggest one. I waited until I had "interesting" numbers to share. But the build in public movement isn't about showing off when you win. It's about being honest while you're figuring things out. The transparency is what builds the audience, not the revenue.
A Genuine Recommendation (Not a Forced CTA)
Okay, here's the part where I have to talk about the Global API affiliate program one more time, but I want to be real about why I'm recommending it.
I'm recommending it because it's the program that's actually paying me every single month. I've got customers I've referred through their affiliate link who are still in month five of their subscription, and I got paid again this month on every single one of them. That's the test. Programs pay out a launch bonus and then quietly change terms. Programs pay out month one and then have weird clawback policies. The fact that Global API keeps paying me, month after month, on customers I referred months ago, is the only endorsement that matters.
Here's the structure, which I want to be specific about because I hate vague affiliate claims:
- 15% commission on first orders when someone signs up through your affiliate link
- 8% recurring commission on every renewal after that
- 10% premium rate available on higher tiers If you send them a customer who pays $300/month, you make $45 on month one and $24 every month after. Send them ten of those customers and you've got $240/month in passive income from a single platform. That's real money, and you don't have to handle support, billing, or infrastructure for any of it. If you've been reading this whole breakdown and you've been thinking "I should try this," I'm going to genuinely encourage you to check out the affiliate program at https://global-apis.com/affiliate. It's how I got started. It's how I still earn a meaningful slice of my monthly revenue. And based on my experience, the payouts actually show up. --- # # Final Thoughts from Month Six I want to end this honestly. I'm not making six figures yet. I'm not quitting my "real job" because this is my real job now — I quit my real job six months ago to do this. Some months are better than others. Some customers leave. Some months feel like I'm pushing a boulder uphill. But the trajectory is real. The compounding is real. The freedom of waking up and checking a dashboard that has revenue in it — even if it's not yet the number I want it to be — is real. If you're reading this and you're in month zero, just know that month zero to month one is the hardest jump you'll ever make. After that, the muscle is built. After that, you have customers. After that, you have
Top comments (0)