Last year I was doing audio for PBS. Twenty years in production, most recently mixing live broadcast content for public television. Then I was laid off at the same time I was diagnosed with Rheumatoid Arthritis. I ended up walking with a cane on and off. Couldn't find work in my field — studios were already downsizing across the board, and I wasn't going to out-compete younger engineers who could physically do more and charge less.
I had to figure something else out.
I started talking to Claude. A lot. I didn't know what I was doing or where it would lead — I just knew I needed to learn something new fast, and AI was the most powerful tool I'd ever touched. Through those conversations, a good bit of deep research, and absolute mountains of trial and error, I stumbled into fintech compliance. Specifically: the gap between what enterprise compliance tools cost ($200-500/month) and what a small developer can actually access.
I decided to build in that gap.
I had no coding background. What I had was time, curiosity, and a willingness to fail publicly. My education was failure — figuring out the hard way what I could and couldn't build with AI as my primary teacher.
That process taught me something I didn't expect: in AI-assisted development, the best lessons come from doing exactly what the AI tells you is impossible.
I started building agents before I knew what an agent was, because I knew I'd need that level of capability to operate alone. I built recursive backend systems to run things I couldn't run manually. The results surprised me. I'm moving faster than I have any right to — not because of prior knowledge, but because of consistency and a refusal to accept "you can't do that" as a final answer.
What I actually built:
A global sanctions screening API. One POST request fans out simultaneously across seven official government databases and returns a unified result:
OFAC SDN — US Treasury — ~18,700 entries
UK OFSI — HM Treasury — ~5,100 entities
EU Financial Sanctions FSF — ~5,800 entries
UN Security Council 1267/2253 — ~1,000 entries
US Consolidated Screening List — 11 export control sub-lists — ~25,400 entries
Canada SEMA — ~5,200 entries
Australia DFAT — ~3,700 entries
Total: ~65,000 restricted parties across every major Five Eyes jurisdiction plus EU and UN. One integration. One response. All refreshed automatically from official government sources.
It's live on RapidAPI. PRO plan is $29/month. Free tier gives you 100 requests to try it without a credit card.
I also have individual screeners for each list at around $15/month. When I find ways to make a screener cheaper, I reduce the monthly price. When I find more content to add, I add it for free.
The honest part:
Zero customers yet. This is a Show IH post, not a success story. I built the thing, the infrastructure is solid, and now I'm looking for the first person who needs it enough to pay for it.
If you're building in fintech, payments, or international business and need to demonstrate sanctions compliance — I'd love to talk. The free tier is genuinely useful for evaluation. If you write about compliance or developer tooling, I'm happy to set you up with access, no strings.
The listing: https://rapidapi.com/Gorp405/api/global-sanctions-export-control-screener
And if you're somewhere in the middle of your own "I have no idea what I'm doing but I'm doing it anyway" phase — keep going. A few months ago I couldn't have told you what an API was. Now I'm shipping products that compliance teams at fintechs actually need. The gap between where you are and where you want to be is mostly just repetition.
Happy to answer anything about the build, the data sources, or what it's like to learn this stuff from scratch.
P.S. - This community is INCREDIBLE. Turns out you don't have to dive into this stuff alone. Wish I would've known about IH months ago. If you're struggling with a build or just the loneliness of starting something by yourself, reach out. I'm happy to chat.