Free tool

Words to Minutes Calculator for English and Chinese

Paste a script and get a reading estimate that counts Chinese by characters and English by words, because one rate cannot serve two writing systems.

AhaNow tools · Updated 2026-11-03 · No sign-up, nothing leaves the page

Summary

This calculator counts Han characters and Latin words separately, applies a speaking rate to each, and adds the two durations. Mixing them into one “word count” is what makes bilingual estimates useless: a Chinese character and an English word are not the same amount of speech, and the ratio between them depends on the sentence. The defaults are starting points you are expected to replace with a measurement of your own voice.

Estimate a bilingual script

Han characters are counted one each; runs of Latin letters and digits are counted as one word each. Nothing else is counted.

Paste the spoken text. Headings and production notes only make the estimate longer.

Words per minute. Measure yours before you schedule against it.

words/min

Han characters per minute, a different unit from an English word.

chars/min

Estimated reading time

0:09

Han characters14
Latin words14
Chinese part3.5 s
English part5.6 s
Total seconds9.1

Punctuation is not counted and pauses are not timed — those are the same thing said twice. Numbers count as one token and are read as several.

Two units, two rates, one total. A Chinese character and an English word are not interchangeable, and no conversion factor makes them so.

What the calculator works out

It counts two things in the text you paste: Han characters, one each, and runs of Latin letters or digits, one per run. Each count is divided by its own speaking rate, the two durations are added, and the total is shown alongside the split so you can see which half dominates. Punctuation, spaces and every script other than those two are skipped.

The split matters more than the total in one specific case: deciding whether a bilingual script is balanced. If the estimate says two minutes of Chinese and two minutes of English, a reader who is fluent in one language and slower in the other will experience a very unbalanced four minutes, and no single total would have shown that.

How to fill in the three fields

  1. Paste the words that will be spoken. Headings, speaker labels and production notes are not read aloud, and every one of them inflates the estimate.
  2. Set the English rate to your own if you have measured one. If you have not, leave the default and remember that the result is a convention, not a measurement.
  3. Set the Chinese rate separately. The two rates are not derived from each other — a speaker can be slow in one language and fast in the other, and the same person is often slower in a second language.
  4. If the script is mostly one language with a few foreign names, do not correct those by hand. The counter already adds them to the other side of the split, and the effect is small.

Why characters and words are different units

An English word is a unit of meaning written with several letters; a Han character is a unit of meaning often written as a single glyph, and many words in Chinese are two characters long. Counting both as “words” means choosing a conversion factor, and that factor changes with the text: a formal announcement and a chat message have different character-per-word ratios. Counting each script in its own natural unit avoids inventing the factor at all.

The practical consequence is that the two rates are not comparable numbers and should never be averaged. 240 characters per minute and 150 words per minute describe the same narrator only in the sense that both are about the same amount of speaking time per minute — which is why dividing one by the other to get a conversion factor is circular reasoning.

When the estimate is wrong

The count is exact and the time is a convention. Every error here comes from the rate being a default rather than a measurement, or from the text containing something the counter cannot see.

  • Numbers, acronyms and names are read longer than their token count. “NASA” is one token and four syllables; a date is one token and a phrase.
  • Pauses are not counted, and punctuation is where they would be. A script written for a slow, deliberate delivery estimates faster than it plays.
  • Text in kana, Hangul, Thai or Arabic is invisible to the counter, so a mixed-script script comes out short by however much of it is in those scripts.
  • Treating the estimate as a duration is the most common misuse. It is a comparison: two drafts, same rates, which one is longer, and by how much.

Getting a number you can schedule against

Record one minute of the actual script at the actual pace, in each language it uses. Count what you got through, put those two rates back into the fields above, and re-read the total. That replaces a convention with a measurement and takes about five minutes — less time than re-recording a segment that ran long.

Questions

Does the counter include punctuation?

No. Punctuation is neither a Han character nor a Latin word, so it is skipped — which is correct for counting, and wrong for timing, because punctuation is exactly where the pauses live. If a script is dense with dashes and ellipses, the estimate will be short.

Why does my word processor count more English words than this page?

The counter splits on anything that is not a letter or a digit, so “well-known” counts as two words and “don't” as two as well. Word processors usually keep hyphenated compounds and contractions together. Neither is wrong; they are different definitions of a word, and the rate has to match the definition you use.

What Chinese rate should I use?

Measure it. The 240 characters per minute default is a starting point, not a measurement: read one minute of finished Mandarin narration aloud, count the characters you got through, and use that. News reading and audiobook reading sit at noticeably different paces, and only you know which one you are making.

Can it estimate a script that mixes Chinese and English?

Yes, that is the reason the tool counts two units instead of one. Each part gets its own rate and the durations are added, so a script that is mostly Chinese with English product names is estimated by adding a short English duration to a long Chinese one.

What happens to Japanese or Korean text?

It is not counted. Kana and Hangul are neither Han characters nor space-separated Latin words, and forcing them into either counter would produce a confident-looking wrong number. Paste that text into a tool built for the language instead; this one will under-report it.

How are numbers counted?

As a single token. “2026” counts as one Latin token here but is spoken as two or three words, and the same is true of years, prices and phone numbers in either language. Numbers are the most reliable way to make an estimate come out short, so read those lines out loud once.