Skip to content
LanguageHelper

句子拆分

把密集的段落拆成每行一句,方便翻译、制作字幕、构建 NLP 数据集或校对。

选项

0

结果
下载

 

100% 在浏览器中运行

关于此工具

Splitting on every period breaks on “Dr. Smith”, “e.g.”, “U.S.A.” or the German “z.B.”. This tool uses the browser’s Unicode UAX #29 sentence segmenter (Intl.Segmenter) for the selected language and then repairs splits after known abbreviations, initials and German ordinal numbers. Chinese and Japanese full-width terminators, the Hindi danda and Arabic question marks are recognized, and Thai — which has no sentence punctuation — is split at spaces. Optional numbering and blank lines make the output ready for spreadsheets, CAT tools or subtitle editors.

与 LanguageHelper 上的所有工具一样,句子拆分 完全在你的浏览器中运行。文本绝不会被上传、存储或用于训练,因此可放心处理机密文件、客户数据和私人消息。无需注册或安装,加载后即使离线也能继续使用。

使用方法

  1. Select the language for the right abbreviation list.
  2. Paste a paragraph or a whole document.
  3. Toggle numbering or blank lines between sentences.
  4. Copy or download the list.

使用场景

  • Prepare source text for translation memory and CAT tools.
  • Create one-sentence-per-line input for subtitles and captions.
  • Build NLP and machine-learning datasets.
  • Proofread long paragraphs sentence by sentence.

示例

English

处理前

Dr. Smith arrived at 5 p.m. today. He was late! Why?

处理后

Dr. Smith arrived at 5 p.m. today.
He was late!
Why?
Chinese

处理前

我很好。你呢?谢谢!

处理后

我很好。
你呢?
谢谢!

使用技巧

Pick the text’s language so its abbreviation list is used — “z.B.” is only an abbreviation in German. If a split happens after an unusual abbreviation, join the two lines by hand. For text without punctuation, run the AI Punctuation Restorer first, then split.

常见问题

How does it handle abbreviations?

Each language has a list of common abbreviations (Mr., Dr., etc., z.B., p.ej., т.е.) plus detection of initials and dotted acronyms. Splits after these are merged back automatically.

Does it preserve paragraphs?

Paragraph breaks are respected: sentences never merge across blank lines. Line breaks inside a paragraph are treated as spaces.

Which browsers are supported?

All modern browsers support Intl.Segmenter. On older browsers a regular-expression fallback is used automatically.

Can I use it for subtitles or translation memory?

Yes. One sentence per line is the standard input format for translation tools, alignment and subtitle editors.

Is the text uploaded?

No, segmentation runs instantly in your browser.

What happens with quotes and brackets?

Closing quotation marks and brackets that follow a sentence terminator stay attached to the sentence they close, so “He said “Yes.” Then he left.” is split after the closing quote, not before it.

Can I split text without punctuation?

A splitter needs sentence-ending punctuation to work with. For raw transcripts, first run the AI Punctuation Restorer, which adds periods and question marks, then split the result here.