Generating multiple languages from a single project

Direct answer

A project's target languages field can hold more than one locale, so a single project generates listing copy for every selected language from the same description, category, tone, and competitor inputs. This keeps every language version anchored to one source of truth about what the app actually does, instead of each locale being generated in isolation and quietly drifting in claims or emphasis.

Why shared inputs matter more for language than for anything else

Screenshot style is cosmetic and platform fields are mostly mechanical, but language output is where drift is most damaging — two locales that describe the same app differently are exactly the kind of inconsistency reviewers and users notice first. Generating every target language from one shared input set is the main structural defense against that, because every locale starts from the same facts.

What a shared source still doesn't fix

Sharing inputs prevents factual drift, but it does not replace review. Idiom, tone, and cultural fit still need language-specific judgment — a phrase that reads as confident in English can read as arrogant in another language even when the underlying claim is identical. Generation from one source is the input-side half of the job; a locale-by-locale review pass is the other half.

Adding a language after launch

When a new target language is added to an existing project, generation reads from the same description, category, and competitor fields already on file — it does not require re-entering the app's story from scratch. That's the practical benefit of keeping languages inside one project instead of starting a fresh project per locale.

Common mistakes

• Starting a separate project per language "to keep things clean," which is exactly the setup that makes claims drift across locales unnoticed.
• Treating shared generation as a substitute for a native or fluent reviewer checking each language output.

Checklist

1. Add every target language to the same project rather than spinning up parallel projects.
2. Re-check the shared description and competitor fields whenever the app's positioning changes, since every locale reads from them.
3. Route every generated locale through a language-specific review pass before it ships — generation is not the same as fluency review.

Operating rule

One shared input set per app, one review pass per language — the first prevents factual drift, the second catches everything generation can't judge.