Why old translation tools don’t fit modern products

Translation tools were invented in the 1990s to manage files for human translators. They assumed translation happens in batches: export your text, send it out, import the results, repeat. That approach only made life easier for one person: the translator.

That worked when software shipped a few times a year. It breaks down when products ship every day, design and code move together, and AI can draft a translation in seconds.

From a place you send files to a tool that’s always on

A better answer isn’t a nicer version of the old tool. Instead of a place you export files to, translation should be always on: connected to your product, noticing when something changes, and keeping every translation up to date.

It should know what text exists, what it says, and what it should say in every language. And it should know when that text changes, when a translation breaks, and when the wording drifts from your brand.

A good translation is more than correct words

A good translation isn’t only correct. It also fits the layout, keeps any names or numbers that get filled in later, matches your brand’s tone, and stays accessible. You should judge it in the product, not in a spreadsheet.

That’s the idea behind StringsOS: translate the whole product experience, not just the words.