I Almost Paid for a Gender API. The Data Was Free.

TL;DR — I built Gender Engine, a free, fast, open-source API that finds the likely gender behind a name. It runs on a public-domain dataset, so you can use it at genderengine.kianreiling.com or self-host it. There’s a live terminal further down — go play with it.

How it started

Initially this problem had a really simple motivation. At work we were thinking about sending out emails with gendered salutations. But because our leads don’t give us any information about their gender, we started thinking about ways to extract the gender from whatever information we did have. The simplest solution would be to use one of the many gender-name API providers out there.

We quickly threw the idea out: gendered salutations have virtually no benefit over gender-neutral ones, but getting the gender wrong in a salutation is very bad. So — high risk, no reward.

The realization

While exploring these different tools, I noticed they all looked very similar. And in the end it always has to come down to some form of dataset — one that reveals that out of 100 people named Sarah, 100 are female.

After a little googling, I found exactly what I was looking for: the WGND dataset, the World Gender Name Dictionary. It contains millions of observed name-country-gender pairs. Exactly what I needed. And to my surprise, the dataset is in the public domain, which means it’s free for everyone to use in any scenario.

So why would I pay for an API that basically just wraps a dataset that looks very close to the one I’d just found?

Just to be clear: these tools probably have good reasons to exist. There are people behind them who work hard to maintain and improve them, and I don’t claim that they all use this dataset.

Learning Go along the way

Anyway, I decided this would be a great project to learn some Go and fiddle around with different storage strategies. If you want to read about how I landed on SQLite as my database and how I came to that decision, you can read about it here (article coming soon).

Where Claude came in

I had written most of the code without AI help, and I’d say I got roughly 70% of the way there. But finishing these projects — deploying them, writing proper tests, publishing them — took actual years. Going over the finish line always seemed too resource intensive.

With Claude Code, that changed immediately. I just wanted to try the technology out and tinkered around at first, but very quickly I realised that Claude Code put me over the finish line — and it did so in a few evenings. It felt exhilarating to finish multiple long-put-off projects in such a short time. Even though I’ve been programming with AI for a long time, the jump in speed and accuracy with Claude Code and Opus is astonishing.

Try it yourself

Essentially this project is a really cheap and fast lookup API that you’re free to use at genderengine.kianreiling.com, or to self-host yourself. Type a name below and hit Enter (try help while you’re at it):