본문으로 건너뛰기
DevTools / 텍스트 통계 · 표기법 변환

텍스트 통계 · 표기법 변환

Word counts matter for far more than essays: a meta description that exceeds its limit is truncated in search results, a commit message over the conventional length is rejected by some linters, and a string field with a byte limit cares about UTF-8 length rather than character count. This tool reports all of those together with sentence and paragraph counts, reading and speaking time, a stop-word-filtered frequency chart, and a case converter that handles the naming conventions code actually uses.

전적으로 브라우저에서 실행 — 아무것도 업로드되지 않습니다.

Characters
263
Without spaces
220
Words
63
Unique words
36
Sentences
4
Paragraphs
2
Lines
5
Non-empty lines
4
Letters
210
Digits
0
Punctuation
10
Whitespace
43
CJK characters
20
UTF-8 bytes
309
Reading time
17s
@ 220 wpm
Speaking time
29s
@ 130 wpm
Word frequency — 30 distinct
  • nothing3
  • is2
  • anywhere.1
  • browser.1
  • customer1
  • data1
  • developer1
  • drop1
  • entirely1
  • every1

Common stop words are excluded from the frequency table.

Case conversion

Case buttons rewrite the text above in place — use Clear to start over.

동작 원리

Character count and byte count are different numbers and conflating them causes real bugs. A JavaScript string is measured in UTF-16 code units, a database column with a length limit may count characters, and an HTTP header or a fixed-size field counts bytes. A Chinese character is one character but three bytes in UTF-8; an emoji is often two code units and four bytes. This tool reports characters, UTF-8 bytes and code points separately, so a value that fits in the interface and overflows the column is visible before it is written.

Word counting is language-dependent and there is no single correct definition. For space-delimited languages a split on whitespace is close enough, with an adjustment for punctuation attached to words. For Chinese and Japanese, which do not delimit words with spaces, a whitespace split produces a count of one for an entire paragraph — so CJK characters are counted individually and reported separately. The reading-time estimate of 220 words per minute and the speaking estimate of 130 are the conventional figures, and both are approximations that vary by roughly a factor of two between readers.

The frequency chart tokenises the text, lowercases it, strips punctuation, removes a stop-word list and then counts. Without the stop-word filter the chart is useless: the, of and and dominate every English text. With it, the words that carry meaning surface. The chart is what a writer uses to find accidental repetition and what a developer uses to sanity-check a tokeniser. Distinct word count and the ratio of distinct to total words — the type-token ratio — are reported alongside as a rough measure of vocabulary variety.

The case converter is the part developers reach for most. It converts to camelCase, PascalCase, snake_case, SCREAMING_SNAKE_CASE, kebab-case, dot.case, Title Case, Sentence case and lowercase, and it splits on the boundaries that actually appear in real identifiers: spaces, hyphens, underscores, dots, slashes, and the transitions between lower and upper case. That last rule is what makes an existing camelCase identifier convert correctly into snake_case rather than arriving as one unbroken word.

주요 활용 사례

  • Check a meta description or title against the character limits search engines display.
  • Measure the UTF-8 byte length of a string before writing it into a fixed-size database column.
  • Estimate reading time for documentation or speaking time for a script.
  • Find overused words in an article before publishing it.
  • Convert a field name between camelCase, snake_case and kebab-case when moving between languages and conventions.
  • Confirm that a JSON field name will survive a transformation that changes case.

최종 수정: 2026-10-14 — 본문은 영어로 표시되며 UI는 번역되었습니다.

자주 묻는 질문

Why is the character count different from the byte count?

Because UTF-8 uses between one and four bytes per character. ASCII letters, digits and punctuation take one byte each, most Latin accented characters take two, CJK characters take three and emoji take four. Any system with a byte limit — an SMS segment, an HTTP header, a database column defined as VARCHAR with a byte length — will reject text that a character count says fits. When a value is rejected for being too long despite looking short, the byte count is the number that matters.

How are Chinese and Japanese texts counted?

Character by character, because there are no spaces to split on. A whitespace-based word count returns 1 for an entire Chinese paragraph, which is technically consistent and practically useless, so CJK characters are counted as individual units and reported as a separate figure. The reading-time estimate uses a slower rate for CJK text, since the conventional words-per-minute figure is calibrated for space-delimited languages and is badly wrong for Chinese.

What counts as a sentence?

A run of text terminated by a full stop, question mark, exclamation mark or their full-width equivalents, with abbreviations and decimal numbers excluded so that e.g., Dr. and 3.14 do not each end a sentence. The heuristic is good on ordinary prose and unreliable on code, tables of numbers and text without terminal punctuation. Treat the figure as an estimate, and note that it is genuinely useful for prose even if it is not a parser.

Which case convention should I use for a variable name?

Whichever your language's community expects, since consistency matters more than the choice. camelCase for variables and functions in JavaScript, Java and Swift; PascalCase for types and exported components; snake_case for Python, Ruby and Rust variables; SCREAMING_SNAKE_CASE for constants in most languages; kebab-case for CSS classes, HTML attributes and URLs, where case-insensitivity makes it the only safe option. The converter will not decide for you, but it will make the mechanical part of switching painless.

관련 도구