💡 Adding Vietnamese (culture code
vi) to a UE5 game takes about half a day of technical work: open Tools > Localization Dashboard, addviunder Cultures, gather strings marked withNSLOCTEXT, export PO files, translate, reimport, and compile to LocRes. The two gotchas unique to Vietnamese are font coverage for its 134-character diacritic set and Unicode NFC normalization - both are straightforward once you know to look for them.
Key takeaways
- Open Tools > Localization Dashboard, add culture code
viunder Cultures, and set a Native culture (usuallyen). - Mark every player-facing string with
NSLOCTEXT("NS","Key","Source")in C++, or use a String Table asset in Blueprints - never plainFString. - Export PO files, translate with Poedit or a CAT tool, reimport, then click Compile to generate LocRes before packaging.
- Use a font that covers Latin Extended Additional (U+1E00-U+1EFF): Noto Sans, Source Sans 3, or Roboto all work in UE5.
- Normalize all Vietnamese text from user input to NFC before displaying or comparing - different input methods produce different Unicode sequences for the same character.
Vietnam had 79.8 million internet users as of early 2025, with over 127 million active mobile connections across roughly 100 million people (source: DataReportal Digital 2025 Vietnam). Gaming ranks among the top mobile categories by time spent. Most titles entering Vietnam arrive in English or Simplified Chinese - Vietnamese localization is a low-cost differentiator in a fast-growing market. The language has no grammatical gender and no plural inflections, reducing translation time versus inflected European languages.
Opening the Localization Dashboard and adding Vietnamese
In the UE5 Editor, open Tools > Localization Dashboard. If you do not have a localization target yet, click Add New Target and name it (usually "Game"). Under the Cultures tab, click + and type vi. This is the BCP 47 language tag for Vietnamese. The regional subtag vi-VN is technically valid but adds no practical benefit unless you plan region-specific builds for overseas Vietnamese communities. Set your source language as the Native culture.
Under Gather, point the gatherer at your source code and content folders. Run Gather Text. This sweeps all NSLOCTEXT macros, Blueprint text pins, and String Tables into the manifest. Then use Export to produce a .po file for the vi culture and send it to your translator.
Once translations are back, use Import to load the PO back into the archive. Run Compile to generate the .locres and .locmeta binary files the runtime reads. Enable vi in Project Settings > Packaging > Localizations to Package before your final build. The ICU data set adds roughly 1.8-15 MB to the build depending on how many locales you support.
Which strings need NSLOCTEXT, and which can stay as FString?
Any text a player reads in-game belongs in an FText created with NSLOCTEXT or LOCTEXT. This covers button labels, NPC dialogue, tutorial text, error messages, tooltip descriptions, achievement names, and quest objectives. String Tables are a data-driven alternative: create an ST_Game.csv asset and reference rows with LOCTABLE. Blueprint text pins wrap FText automatically and are gathered without extra work.
Strings that should not be localized include debug log messages, server API keys, analytics event names, asset reference paths, and any string that is never displayed to the player. Use INVTEXT or plain FString for those. Accidentally marking an internal identifier with NSLOCTEXT only bloats the manifest and creates noise for translators reviewing thousands of strings.
FText formatting: numbers, currency, and argument modifiers
Use FText::Format() with named arguments for strings that change at runtime. Vietnamese Dong (VND) has no sub-units, so format currency with zero decimal places:
FText Msg = FText::Format(
NSLOCTEXT("HUD","GoldMsg","Gold: {Amount}"),
FFormatNamedArguments{{"Amount",
FText::AsCurrencyBase(PlayerGold, TEXT("VND"))
}}
);
FText::AsCurrencyBase() uses ICU culture rules: for vi, it applies periods as thousands separators and no decimal places. FText::AsDate() outputs day-first order (dd/MM/yyyy) as expected by Vietnamese readers. The plural modifier {Count}|plural(one=item,other=items) is available in Unreal, but Vietnamese has no grammatical plural. A good translator will restructure the sentence to avoid plural forms entirely rather than using the modifier. Vietnamese uses noun classifiers (cái, con, chiếc) instead of inflection, and those vary by noun category.
Does Vietnamese text run longer than English in your UI?
Vietnamese strings are often close in word count to English, but each character carries stacked diacritics. A single character like ộ (o + circumflex + dot below) renders taller than a plain ASCII letter. This means Vietnamese text needs more vertical line height: around 10-20% more than Latin text without diacritics. Buttons sized for English labels can clip Vietnamese text at the top and bottom rather than on the sides.
Test your most constrained UI layouts with real Vietnamese strings early. A quick proxy: add "g" or "y" to the end of English words to approximate vertical overflow. For line length, Vietnamese tends to stay similar to English because shorter Vietnamese words compensate for the diacritic overhead. The UI risk in Vietnamese is vertical, not horizontal. Plan extra line-height in your UMG text styles before final layout lock.
Which fonts cover the full Vietnamese diacritic set in UE5?
Vietnamese characters span several Unicode blocks: basic Latin (U+0041-U+007A), Latin Extended Additional (U+1E00-U+1EFF), and Combining Diacritical Marks (U+0300-U+036F). A font must include the full Latin Extended Additional block to render all tone marks and vowel combinations. Missing glyphs appear as boxes or question marks during play.
Well-tested options that import cleanly into UE5 include Noto Sans, Source Sans 3, and Roboto. Import the .ttf as a Font Face asset, create a Font asset, and assign it to your Slate or UMG text styles. If your existing game font lacks Vietnamese coverage, set up a Composite Font in UE5: add the Vietnamese-capable font as a fallback range for U+1E00-U+1EFF. The engine then blends the two fonts per character - your brand font handles basic Latin while the fallback covers Vietnamese diacritics. Test by forcing the culture to vi and scanning every UI screen for boxes.
Unicode normalization: why NFC is mandatory for Vietnamese
Vietnamese keyboards and input methods produce different Unicode sequences for the same visible character. The word "phở" can be encoded as a single precomposed codepoint or as the base character "o" plus two combining marks. The W3C Unicode migration guidance explicitly names Vietnamese as a language where input normalization is required, noting that Windows input methods and third-party Vietnamese software produce incompatible codepoint sequences for the same word.
Normalize all Vietnamese text from user input, external APIs, or file I/O to Unicode NFC before storing, displaying, or comparing. Without this, two players who type the same Vietnamese username may produce strings that compare as unequal - causing bugs in search, friend lists, achievement tracking, and any feature that compares text. The ICU library that UE5 ships handles culture-aware collation, but the input must be normalized before you hand it to ICU.
What your studio can handle in-house vs when a Vietnamese specialist is worth it
Your technical team can handle the full pipeline in this guide without outside help: Dashboard setup, PO export and import, font selection, NFC normalization, and testing with -culture=vi. Machine translation tools such as DeepL and Google Translate produce usable first drafts for simple UI strings - button labels, menu items, and system messages with minimal context.
A native Vietnamese specialist is worth bringing in for:
- Game narrative and dialogue: machine-translated NPC speech sounds flat or awkward and breaks immersion immediately for Vietnamese players.
- App store metadata: title, short description, and long description are the first impression and directly affect discoverability on Google Play and the Apple App Store in Vietnam.
- Culturally specific names and jokes: Vietnamese players notice when a character name has an unintended meaning, or when a joke uses the wrong register for the target audience.
- North vs South register: standard written Vietnamese is universal, but word choices like "xe may" (South) vs "xe gan may" (North) signal origin to native readers.
A review pass by a native Vietnamese game localization specialist after a machine-translation first pass catches most of these issues at a fraction of full translation cost.
What culture code should I use for Vietnamese in Unreal Engine?
Use vi as the BCP 47 language tag. The regional subtag vi-VN (Vietnam region) is technically valid but adds no practical difference unless you plan region-specific builds for overseas Vietnamese communities such as Vietnamese-Americans.
Can DeepL or Google Translate generate my Vietnamese PO files automatically?
Machine translation handles short, context-clear strings such as button labels and menu items reasonably well. It struggles with longer dialogue, idiomatic expressions, and culturally specific content. Use MT as a first draft for UI strings, then have a native Vietnamese speaker review dialogue, marketing text, and store metadata before shipping.
Do North and South Vietnamese players need separate localizations?
No. Standard written Vietnamese is mutually understood across all regions. Spoken dialects differ in vocabulary and pronunciation, but published games use standard written form. The only exception is voiceover: both a Northern (Hanoi) and a Southern (Ho Chi Minh City) accent are acceptable for voiceover, but picking one may affect perception with the other regional audience.
How do I test Vietnamese localization without a Vietnamese QA team?
Run the game with the culture forced to vi using the command-line flag -culture=vi, or change the preview culture in the Localization Dashboard. Check for clipped text, missing font glyphs displayed as boxes, and layout overflow. Have a native speaker review a sample of strings - especially dialogue and store metadata - before full QA.
How should I handle Vietnamese text entered by players, such as usernames or chat?
Normalize all player-entered Vietnamese text to Unicode NFC before storing or displaying it. Vietnamese input methods produce different codepoint sequences for the same character depending on the method used. Without normalization, two players who type the same username may produce strings that compare as unequal in search or matchmaking features.
Official Sources
- Unreal Engine Localization Overview - file formats (manifest, archive, PO, LocRes, LocMeta), workflow, ICU library. Verified November 2026.
- Unreal Engine Text Localization - FText, NSLOCTEXT, format modifiers, String Tables, FPolyglotTextData. Verified November 2026.
- W3C: Unicode Migration - Vietnamese Normalization - Vietnamese input normalization and NFC requirement. Verified November 2026.
- DataReportal: Digital 2025 Vietnam - 79.8 million internet users, 127 million mobile connections. Verified November 2026.

Written by Dao Huy (Lucas), Vietnamese translator and localization specialist (English, Chinese and French into Vietnamese). See translation services →
