A character-based conversion of text into another writing system to approximate its pronunciation without translating meaning.
Transliteration replaces characters from one script with characters from another script, rather than translating meaning. The goal is to make words readable to users who do not know the original writing system but can recognize the sounds.
In localization, transliteration is commonly used for proper names, place names, brand names, and user input. It allows content written in scripts such as Cyrillic, Arabic, Greek, or Devanagari to appear in Latin characters, or the other way around, without changing the original word itself.
Unlike translation, transliteration does not adapt meaning. And unlike phonetic transcription, it usually follows spelling conventions rather than precise spoken sounds. Because scripts do not always map cleanly to each other, transliteration often follows established standards or language-specific rules.
Different transliteration systems can produce different results for the same word. Some sounds do not exist in the target script, which leads to approximations. Inconsistent transliteration can confuse users or fragment search results, so teams usually define clear rules and apply them consistently.
In software localization, transliteration often appears alongside translation rather than replacing it. A product may translate interface text but transliterate names or user-generated content to preserve identity and clarity.