In previous posts I wrote about building a browser-based board game called "Growing City" in three days with AI, and later about spending a couple of weeks polishing the lobby, bots, and UI. This time it's about a decision that might help a lot of people building products for a global market.
A Game That Doesn't Care Who Plays It
An economic city-building board game isn't tied to any region. The rules are the same whether you're in Moscow, Madrid, or New York: roll the die, build something, outplay your opponents. Nothing specific that would only work for one audience.
Yet the site was only in one language. Not by choice — when you ship an MVP in three days, you don't think about a three-language interface, you think about making it work at all. But it felt wrong to artificially cut off my audience with a language barrier that didn't even exist in the game mechanics themselves.
Translation Isn't the Hard Part. It's Literally the Easy Part
Here's what I did technically, in short: all interface text got moved into separate dictionary files — one for Russian, one for English, one for Spanish — each with the same set of keys. The "Buy" button isn't text hardcoded into the code; it's a key that gets swapped based on the selected language. I also had to separate card names into two categories: what the player sees (translatable) and what the game engine uses internally for logic (not translatable, must never change). Otherwise, translating the interface would have quietly broken the game itself.
The neat part was how I handled the default language. No need to ask the user anything on first visit — the browser already knows what language the user's system is set to. Someone visits from a Spanish OS, the interface is instantly in Spanish, no extra click needed. And if they manually switch the language later, that choice is remembered and never gets overwritten by the automatic detection again.
The Most Unexpected Part: Not the Translation, But Everything Around It
Writing an article isn't a problem. Translating it into English and Spanish isn't a problem either — that's literally one prompt. But then comes what I already wrote about in my first post: the difficulty shifts from code to everything surrounding the product.
Posting an article to a dozen different platforms for different audiences used to mean manually opening every single site, pasting the text, adjusting the headline to fit each community's format. It was eating me alive — pure repetitive busywork that shouldn't need to exist.
This is where an AI browser agent came in handy — a tool that can see the browser page and click, fill out forms, and publish content on its own. I just gave it a link to an already-finished article and asked it to post it to public platforms. And that was it — done. Literally one prompt and about 10 minutes, and the article was live on international platforms (just like this one).
What's Next
The game is still live at rastushiy-gorod.ru, now available in Russian, English, and Spanish depending on who visits. You can join without registration and play with friends, against bots, or with a random stranger from the other side of the world who landed in the same open lobby.
International users have already started showing up. It's too early to draw conclusions, but I'll share numbers and early observations in the next post.
As always, I'm collecting feedback and would love to hear your thoughts in the comments.
Top comments (0)