Counting a Korean script: spaces matter

Choose a unit before choosing a rate

A Korean eojeol is a whitespace-separated unit. It is not identical to an English word or a Korean grammatical word. For a practical script estimate, this tool counts tokens containing a letter or number and labels Korean results as eojeol. It also shows the character count without whitespace.

The same speech can produce a different count

Removing spaces from “오늘 영상 만들어요” turns three units into one without making the spoken sentence three times shorter. A rate calibrated on spaced text therefore cannot be applied unchanged to an unspaced script. The tool warns about long unspaced CJK runs, but that warning does not reconstruct the correct word boundaries.

Punctuation, numbers and stage directions

1. A token containing only punctuation or emoji does not count as a spoken unit here.
2. Numbers count as a token, but “2026” can take longer to say than one short word. Rewrite it as spoken text when precision matters.
3. Remove directions like “[show chart]” if they will not be spoken; the tool cannot infer your intent.
4. Line breaks, tabs and repeated spaces separate units but do not add empty units.

A quick spacing check: characters per eojeol

The script tool shows both the eojeol count and the character count, so you can divide one by the other. The Korean sample used on the tool page gives:

Removing the spaces leaves the characters unchanged but collapses the eojeol count, so the ratio jumps. Compare a new script’s ratio with scripts you have already timed. A much higher ratio usually means missing spaces, which would make an eojeol-based estimate far too short. The tool’s warning for long unspaced runs catches the extreme case; the ratio also reveals a script that is only partly unspaced.

Calibrate in the same language

The Korean preset uses the same starting numbers as the English one. It is an editable planning setting, not a Korean reading-rate finding. Use a Korean passage and the timer to measure eojeol per minute. Then apply that rate to similarly spaced Korean scripts. Mixed-language scripts may need separate estimates or an actual read-through.

Characters help spot a spacing anomaly; they do not provide a second guaranteed timing formula. The tool intentionally keeps the counting rule visible so you can decide whether a sample represents your own script.