What it does
The Character Frequency Counter walks through the text once, counts each visible character, and sorts the results from most to least common. Each row shows the character, its count, and its share of all counted characters. The table updates as you type, so you can paste a paragraph, delete a sentence, and watch the ranking change.
The tool counts grapheme clusters, not raw code units. An emoji is one character, not two or four. An accented letter is one character even when it is stored as a base letter plus a combining mark. That matters for languages with diacritics and for any text that contains emoji; a naive count would overstate the length and distort the percentages.
Reading the table
The bar is scaled to the most common character, not to the total. The first row always has a full bar, and the others are relative to it. The percentage column is the share of all counted characters, so the percentages add up to 100% (within rounding) across the rows that are shown.
A long tail is normal. Most English prose has a handful of very common letters and many rare ones. If the table is limited to the top fifteen, the remaining characters are still counted in the totals; they are simply not displayed. Raise the row limit to see more, up to two hundred rows.
Letters and case
“Letters only” is on by default. It filters out digits, punctuation, symbols, and whitespace, leaving a pure letter-frequency profile. Turn it off to see the full character mix, including commas, periods, digits, and currency signs. The switch is useful when you are checking a dataset rather than analyzing prose.
Case is folded by default: A and a are the same character. That is the right choice for language analysis, because capitalization is a stylistic overlay rather than a different letter. Turn case sensitivity on when the distinction matters — for example, when you are studying title case, acronyms, or a puzzle that uses capital letters as a clue.
Practical uses
- Cryptograms and substitution puzzles. The most frequent letter in an English cipher is often E, followed by T, A, O, I, and N. A frequency table gives you a starting hypothesis, not an answer; short texts are unreliable.
- Typography and design. If a display face has a weak Q or an awkward Z, a frequency profile tells you how often the problem will actually appear in the copy.
- Writing analysis. A very high share of one letter can be a stylistic tic, a repeated word, or a sign that the text is not in the language you assumed.
- Data checks. With letters-only off, a table that shows an unexpected delimiter or a high count of a symbol can reveal a formatting problem before a CSV import.
- Language identification. The presence of letters such as ñ, ø, å, or ł is a hint about the language. It is a hint, not proof.
Limits worth knowing
- Short texts are noisy. A single sentence does not follow the frequency profile of a language. The table describes the text in front of you, not English in general.
- Whitespace is excluded from the ranking. Spaces, tabs, and line breaks are not rows. They are still present in the input and are counted in the character total shown in the footer where applicable.
- Grapheme counting is not byte counting. A character may use several bytes in UTF-8, but the table counts visible characters. Do not use it to estimate file size.
- Case folding is locale-aware. Some languages have special casing rules. The tool uses the browser’s locale-aware lowercasing, which is usually right for prose but may not match a database’s collation.
- Nothing is stored. A refresh clears the input and the table. Copy or download the TSV before leaving.
Page updated