Caption vs In-App UI Localization for Store Screenshots

A caption is the headline or tagline you place on the screenshot canvas, usually above or below a device frame. It is marketing copy: short, benefit-led, and written for someone scrolling a store listing.

In-app UI text is everything rendered inside the captured screen: tab labels, buttons, list items, settings rows, onboarding copy and sample data. It comes from your app's string files, not from a design tool, so changing it means recapturing the screen or editing the pixels.

Google Play draws the line most clearly. Its preview asset guidance says: "In-game UI does not need to be localized per market, but any additional taglines or text overlays should be localized" (Play Console Help: Add preview assets (opens in a new tab)). The same page asks you to "Localize your graphic and branding text as appropriate for different markets and languages."

That page also limits how heavy captions can be: "Taglines should not take up more than 20% of the image." Keep that in mind when a translation runs longer than the English source. For video, the bar is higher: Google asks you to localize "the video including the UI, taglines, and audio."

Apple does not split the two levels. If your app is available in multiple languages, Apple says to "localize your app description, keywords, app previews, and screenshots, for each of the markets in which you offer your app" (Apple: Creating your product page (opens in a new tab)). Its localization overview (opens in a new tab) adds that you can "tailor your screenshots to reflect the values of each market."

Caption-only is a reasonable default when the caption, not the UI, does the selling. If a shopper can understand the benefit from the headline alone, translating the headline covers most of the message.

• Games and visual apps where the screen shows gameplay, photos, maps or charts with little readable text. This is the case Google's "In-game UI" wording describes.
• Heavily framed layouts where the device mockup shrinks the UI to the point where labels are not readable on a phone-sized listing.
• Markets you are testing before committing to a full app translation. A localized caption signals intent without claiming the app is translated.
• Apps whose UI is not yet localized. Showing translated UI that users will not see after install sets the wrong expectation.

Localize the UI inside the screenshot when the app ships in that language and the interface text is part of the message. A Japanese caption over an English settings screen reads as a half-finished listing to someone who will use the app in Japanese.

• Text-heavy products: productivity, finance, messaging, education and news apps, where the UI copy is the feature being shown.
• Full-bleed screenshots without a caption, where the UI is the only text on the image.
• Priority markets that already produce meaningful installs, where polish is worth the extra production time.
• Scripts that look different at a glance, such as Japanese, Korean, Chinese, Arabic or Hebrew. A Latin-script UI stands out immediately in those listings.

Plan for layout differences too. The W3C notes (opens in a new tab) that the Japanese word for "desktop" has one character fewer than the English but "will typically take up much more horizontal space." Buttons and tabs that fit in English can truncate in the localized build.

Most teams land on a hybrid: captions everywhere, in-app UI only where it pays off. Apple notes that the first one to three screenshots can appear in search results when no app preview is available (Apple: Creating your product page (opens in a new tab)), so those slides deserve the most care.

1. Translate every caption for every target locale, and check each one against the 20% tagline limit on Google Play.
2. Rank your markets by installs or strategic value, and mark the locales where the app itself ships translated.
3. For those locales, localize the UI on the first one to three slides, where search visibility is highest.
4. Extend UI localization to the full set in your top markets once the first slides are reviewed.
5. Leave visual slides caption-only: maps, photo grids, gameplay and charts rarely need UI changes.
6. Review every locale with a native speaker before upload, including truncation, line breaks and right-to-left direction.

Example: a habit tracker with six slides and ten locales. Captions are translated for all ten. For the four locales where the app is translated, slides one and two get localized UI, and slides three to six keep the source UI because they show charts. That is 8 UI edits instead of 24 for full UI localization in those four locales, with the first slides fully consistent.

Captions are text layers on the canvas, and each text layer stores its own content per locale. Switch the preview language and edit the headline for that market without touching the layout, fonts or colors of the other locales.

• AI text detection and translation cost 0 credits, so translating captions for all 32 target languages is free.
• Translation keeps brand marks and acronyms such as 4K, HD, PRO and AI unchanged, and every string stays editable per locale.
• A side-by-side view shows the source slide next to the localized preview, which makes overflow easy to spot.
• Strict export can skip or block slides that still contain untranslated text.

For text inside the captured screen, StoreLocale uses AI repaint. A vision model detects the UI strings and their positions, the strings are translated, and an image model replaces each one in place while trying to match font, weight, color and kerning. Video players, posters and thumbnails are masked and restored from the original pixels, and rating or format badges such as PG-13 or 4K are left untranslated.

Credits come in one-time packs with no expiry; see pricing. If you use your own OpenRouter or Gemini key, generations use 0 platform credits. Unchanged assets are reused without a new charge.