Translate a JSON i18n File into Simplified Chinese

What changes when your strings become Chinese

Simplified Chinese has no spaces between words and uses full-width punctuation. In a locale file this affects almost every value: "welcome": "Hello, {name}!" becomes "welcome": "你好，{name}！". The comma and exclamation mark are Chinese characters (，！), not Latin ones followed by a space.

• Length: a Chinese value often has two to three times fewer characters than an English or French source, but each character is wider. A button barely gains any room. A field capped with maxlength, or truncation based on character count, will however behave differently.
• Word order: placeholders move. "{count} days ago" becomes "{count} 天前", while "Updated by {user}" becomes "由 {user} 更新". This is why a sentence split across several keys ("You have", then {count}, then "messages") translates badly: each fragment is translated on its own, and your code reassembles them in the source order.
• Measure words: a number needs a classifier. "{count} files" becomes "{count} 个文件", "{count} pages" becomes "{count} 页". The placeholder stays; the classifier is added around it.

Plurals and ICU messages: one category only

According to CLDR, Chinese has a single plural category: other. There are no few or many forms to add, unlike Polish or Russian. An ICU message such as {count, plural, one {# file} other {# files}} keeps its structure and both branches are translated; in Chinese, the other branch is the one displayed, even for 1. That is the branch to review first.

• Exact matches remain separate branches: =0 {no files} becomes =0 {没有文件}.
• select branches are translated one by one, even where spoken Chinese makes no distinction (他 and 她 sound the same).
• With i18next, suffixed keys such as file_one and file_other are neither translated nor renamed. Recent i18next versions follow CLDR rules and will only read file_other for Chinese, so make sure it exists in your source.

For each string, the translation is checked to contain the same placeholders, {name}, {{count}} or %s, as many times as the source.

What Manuce returns, and what it does not do

You get back the same file: indentation, key order, quotes, comments, line endings, encoding and BOM preserved. Only the values change, whether the JSON is flat, nested or contains arrays of strings. Each translated value is compared with its source (placeholders, tags); if a mismatch persists, the source value is kept and the report says so. The report lists the strings translated and those left as they were, with the reason.

Left untouched: keys, empty values, URLs, email addresses, numbers, colours such as #1A73E8, date formats such as YYYY-MM-DD, identifiers and anything you have marked as non-translatable.

• Manuce does not rename your file to zh.json or zh-CN.json, nor change your library configuration.
• It does not turn a date pattern into YYYY年M月D日: date formats are not translated, so Chinese date display belongs in your code (for instance Intl.DateTimeFormat with the zh-CN locale).
• It does not produce Traditional Chinese in the same order: that is a separate language to add (two languages bring 20% off the whole order).

Preparing the file and checking the result

Before uploading:

• Make sure the file is UTF-8. Encoding is preserved, so a file saved in a Western single-byte encoding could not hold Chinese characters.
• Merge split sentences into one key with a placeholder: "inbox": "You have {count} messages" rather than three concatenated keys.
• Mark product names or codes that must not change as non-translatable.

After delivery:

• Read the report: a value left in the source language often points to a malformed placeholder in the original.
• Add a Chinese font to your CSS stack (for example "PingFang SC", "Microsoft YaHei", "Noto Sans SC"): Latin fonts lack these characters and the browser will pick an arbitrary fallback.
• Check line wrapping: with no spaces, long Chinese text breaks between any two characters, which is correct, but fixed-width containers may overflow differently.
• Pick a rule for spacing between Chinese characters and digits or Latin words (共 {count} 条 or 共{count}条) and check it is applied consistently.
• For a brand name or technical term that must stay as it is, look at the free sample before paying, then at the full translation.

In practice

1. Check the JSON is UTF-8 and that each sentence sits in a single key.
2. Upload the file and choose Simplified Chinese: the quote counts words in values only.
3. Read the free sample, ideally a string with a placeholder or plural.
4. Download the translated file and report, and place it under your zh or zh-CN locale.
5. Test the interface with a Chinese font and check length-limited fields.