English word segmenter
Split an English paragraph into standalone words: contractions such as don't stay whole, dotted abbreviations like U.S.A. fold into one token, and hyphenated compounds can be split or kept — with word, unique-word and character counts.
Runs in your browserEvery computation happens in your browser — your data never leaves this device.
Tokens
One item per line, ready to copy.
Paste an English paragraph to segment it
Words: 0 · Unique: 0 · Characters: 0
What this tool does
- Turn a paragraph into a token list for frequency counting, word clouds or vocabulary drills.
- Pre-process text so contractions and hyphens are handled once instead of being mangled downstream.
- Check terminology consistency: see whether compounds such as well-known are written joined or split.
Example
Input
Don't stop; it's a well-known idea. The U.S.A. matters.
Output
Don't stop it's a well known idea The U.S.A matters
don't and it's are single tokens, U.S.A. folds into one token (the final period is dropped), and well-known splits into two words by default.
Frequently asked questions
Is don't split apart?
No — an apostrophe inside a word is part of the token by default, so don't, it's and o'clock each count once. Turning the apostrophe option off splits them.
How is U.S.A. handled?
Inner dots are kept and a trailing period is dropped, giving the single token U.S.A. Turning the dotted-abbreviation option off splits it into U, S and A.
Is a hyphen a separator?
By default yes: well-known becomes well and known. Turn hyphen splitting off to keep it as one word.
Are numbers treated as words?
No. Segmentation only recognises Latin letters, so 123 or 3.14 are dropped. Use the number extractor for digits.
Does the word count include duplicates?
Both numbers are reported: “words” is the total token count and “unique” is the count after folding case.
Keywords：英文分词分词word segmentationtokenizer英文拆分单词切分词数统计U.S.A.