Translate a CSV String Table into Simplified Chinese

What Chinese changes in a string table

A row such as menu_start,Start game becomes 开始游戏 in Chinese: four characters instead of ten letters. Chinese text is two to three times shorter in characters, but each character is wider and taller, so the space taken on screen changes little. A limit counted in characters leaves more room than it seems, while a button sized in pixels stays as tight as it was in English.

• No spaces between words: your renderer must be able to break a line between any two characters. If your code wraps text on spaces, a Chinese sentence will run on as a single line.
• Full-width punctuation: ，。：！？ take up a full character width. The Chinese comma ， is not the CSV delimiter, so it never shifts your columns.
• Ellipsis: "Loading..." becomes something like 加载中…, using the ellipsis character rather than three full stops.
• Quotation marks: Simplified Chinese uses “ ”, which need no escaping in CSV, unlike the straight quote " that must be doubled as "".

Simplified Chinese targets mainland China and Singapore. For Taiwan or Hong Kong, Traditional Chinese is a separate language at Manuce, with its own column.

Placeholders, plurals and word order

Chinese has a single CLDR plural category: other. If your table holds coins_one and coins_other, both rows are translated, usually identically: {0} 枚金币. What Chinese does need is a measure word between the number and the noun: {0} 个文件 (files), {0} 条消息 (messages), {0} 名玩家 (players). A number glued on in code (count + " files") leaves no room for it; a single string with a placeholder works far better.

Word order moves too. Level {0} becomes 第{0}关, with the placeholder in the middle; Reward for {quest} gives {quest}的奖励, with the placeholder first. Named ({player}) or numbered (%1$s) placeholders can move freely; two anonymous %s lock the source order in place and can force awkward Chinese.

Each translated value is compared with its source. If a placeholder or tag ({icon}, [color=red]) goes missing or is altered and the mismatch persists, the source value is kept and the report says so.

What Manuce returns, and what it does not do

Manuce reads the developer-style CSV: one key column (key, id…) and one column per language (en, fr, de_DE…), as used in Unity or Godot. If a Chinese column already exists, it is filled; otherwise it is added as the last column. A comment or context column is used as context: that is what decides between 打开 (open, as a button) and 开放 (open, as in accessible).

The file you get back is the same file: key order, quoting, comments, line endings, encoding and BOM preserved; only the values change. Not translated: keys, empty values, URLs, email addresses, numbers, colours (#FFCC00), date formats (yyyy-MM-dd), identifiers and anything marked as non-translatable. The report lists translated strings and those left as they were, with the reason.

Manuce does not add a Chinese font to your project, configure your engine's line breaking or test the display. It does not take a glossary: for a brand or character name that must stay in Latin letters, check it in the free sample and then in the translation. A CSV with no key column is not treated as a language file but translated as plain text. Pricing counts the words in values only, keys and tags excluded, with the ×1.6 coefficient that applies to Chinese.

Preparing the file

1. Encoding: save the CSV as UTF-8. Manuce keeps the original encoding, and a Windows-1252 or ISO-8859-1 file simply cannot hold Chinese characters.
2. Column name: if your engine expects a specific code (zh_CN, zh_Hans, zh…), create that column empty; it will be filled without being renamed. Otherwise the column added at the end may need renaming.
3. Context: fill in the context column for short, ambiguous strings. "Back" can be 返回 (go back) or 背面 (the back of a card); "Save" can be 保存 (save a file) or 拯救 (rescue someone).
4. Formats: date formats are not translated. A Chinese date reads 2025年3月14日, but your application should pick the format per locale rather than relying on the string.

Checking the result in your project

Inspect the returned file in a text editor rather than a spreadsheet: some spreadsheets misread UTF-8 without a BOM and may change the encoding or quoting when saving. Then, in the game or app:

• Font: a Latin font has no Chinese glyphs; without a fallback, strings show as empty boxes. Plan for a font covering Simplified hanzi, bearing in mind it is much heavier than a Latin one.
• Size: characters with many strokes (警告, "warning") become unreadable at small sizes; tooltip text that works in English may not in Chinese.
• Line breaks: check that long sentences wrap properly and that a ， or 。 never starts a line on its own.
• Report: go through the strings kept in the source language because of a placeholder issue, or they will appear in English in the middle of a Chinese interface.

In practice

1. Save the CSV as UTF-8 with its key column, an empty Chinese column named as your engine expects and, ideally, a context column.
2. Upload the file: the quote appears straight away, based on the exact word count of the values.
3. Read the free sample to check placeholders and proper names, then start the translation.
4. Download the CSV and the report, drop the file back into your project and test font, size and line breaks.