Skip to content


Multilingual Localization Services


Alafranga has been localizing digital content since 2002. Websites, software interfaces, mobile apps, games, database content and subtitles into 40+ other languages.


Most projects still run the classic way. You send us the source files, whether a CMS export, resource files from a software build, game strings or subtitle files, and we translate them in memoQ or Trados with your translation memory and glossary.

For products that change every week, we can build the process instead of repeating it. We connect your CMS, repository or build pipeline to a translation platform such as Smartcat or Phrase, or to an open-source system on your own server, so that new and changed strings are picked up automatically, drafted with AI under your glossary and checked for placeholders, broken tags and length before a specialist reviews them. Nothing ships without that approval.

→ AI Translation Workflow Consulting
App localization operates within strict UI limitations, string logic, and release-cycle dependencies. A single misplaced string can affect build stability, break user flows, or trigger platform compliance issues across every language version simultaneously.

Our software localization workflow is built around your development environment. We handle string extraction, in-context review, pluralization logic, and linguistic validation before release, not after.
 
From string logic to App Store metadata, we validate before release, not after. read more..

▮Web localization services 

Websites are dynamic, responsive, and metadata-driven. Localization here goes beyond translating visible text and it affects SEO performance, compliance visibility, CMS architecture, and conversion behavior across every market you operate in. Our web localization workflow integrates directly with your CMS and live interface constraints, ensuring every language version is as functional and optimized as the source.
 
From metadata SEO to CTA alignment, we catch what automated tools miss. read more..

▮Game Localization Services

Game content combines the constraints of software with the demands of narrative. Interface strings sit inside hard character limits that differ between mobile and desktop. Dialogue has to keep voice and register consistent across hundreds of files and across every update that follows.
 
What we handle: interface and menu strings, dialogue and narrative text, item, skill and achievement descriptions, tutorial and onboarding copy, store listings and marketing text, and LiveOps content on a recurring release cycle.
 
Terminology is locked in a glossary at the start and carried through every update, so a weapon named in the first release is not renamed in the third. We have delivered game localization for studios including HI-REZ Studios and Playata, with Turkish as the core pair. read more..

Accurate subtitling requires both linguistic precision and technical synchronization.
We support:
  • Subtitle translation and adaptation
  • Time-coded synchronization
  • Platform-specific formatting
  • SDH (Subtitles for the Deaf and Hard of Hearing)
  • Multilingual version management
  • Consistency across episodes and seasons ensures brand continuity.
Learn more about media localization→ 

 

▮ Compare Localization Workflows

  • Human translation. Translated and independently reviewed by one or two human specialists, with no machine drafting at any stage. Used for regulated content, legal and privacy text, and anything that contractually cannot be sent to a third-party AI provider. Our human workflow operates to ISO 17100:2015. read more..
  • AI draft with human review. An AI draft is generated inside memoQ or Trados with your translation memory and glossary applied, then reviewed by a human specialist. Faster and cheaper for content that suits it. Conformity to ISO 17100 is not claimed for machine-drafted text. read more..
  • AI draft with two review stages. The same drafting step followed by a second independent review, for content where the risk of a single reviewer missing something is not acceptable. read more..

The workflow does not have to be the same for every language or every content type. AI output quality still varies between languages, so we run a feasibility test per language and content type before applying AI drafting. Our usual starting point:

Content Default workflow What we check
Software and app interface strings Human translation with revision (HumanExpert Pro). AI draft only where the feasibility test passes Length per string, placeholders, plural forms, then in-context review in the running build
Website pages, landing pages, calls to action Human translation with revision (HumanExpert Pro) Title and meta description length, local search terms, calls to action written for the market
Help centre, knowledge base, release notes, developer documentation AI draft with post-editing and revision (SmartEdit Pro) Glossary terms, links, code samples and tags left untouched
App store listings and metadata Human translation with revision (HumanExpert Pro) Character limits per store, keywords in the target language
Game dialogue and narrative text Human translation with revision (HumanExpert Pro) Voice and register across files, character limits per platform
Game items, achievements, LiveOps updates AI draft with post-editing and revision (SmartEdit Pro), if the feasibility test passes Names locked in the glossary from the first release, length limits
Database content, product catalogues AI draft with post-editing (SmartEdit) Volume and reuse from your translation memory, automatic checks on numbers and units
Privacy policies, terms of use, legal text Human translation with revision (HumanExpert Pro), no AI drafting Content that may not go to a third-party AI provider


Interface strings, marketing copy, and narrative dialogue often fail the feasibility test. Help content and documentation usually pass it. We tell you the result for each language before you commit.

▮How a Localization Project Runs

Briefing and asset preparation. Glossaries, style guides, translation memories, string files, screenshots and interface references collected before anyone starts translating.
 
  • Localization. By a linguist with experience in your content type, working in your preferred CAT or TMS environment, inside your string and interface constraints.
  • In-context review. A second linguist checks the content in the running interface, not only in the editor: accuracy, terminology, fit inside the component, and whether it reads like something written for that market.
  • Quality assurance. Automated checks for interface overflow, placeholder integrity, concatenation, formatting tags, numbers and omissions.
  • Delivery and feedback. Returned in your file format, with translation memories and glossaries updated for the next release.