Character Frequency Analyzer (letters, all chars, or words)
Paste any text to get a sorted frequency table. Three modes: all characters, letters/digits only, or whole words. Useful for cryptanalysis, writing analysis, and dataset cleanup.
How it works
What's it for
Cryptanalysis: classical ciphers (Caesar, substitution) preserve letter frequencies. English text always has E as the most common letter, then T, A, O, I, N. If you see roughly that distribution in cipher text, you have a substitution. CJK languages have very different distributions but still recognizable.
Writing analysis: spotting overused words is one of the fastest ways to improve drafts. If 'just' or 'really' appears 50 times in a 1000-word essay, you've found a tic to fix.
Dataset cleanup: scanning a CSV column with this tool reveals stray characters, encoding errors, and unexpected casing. Useful before importing data into a stricter system.
Three modes
All characters: includes spaces, punctuation, line breaks, emoji. Best for raw text analysis. Useful when you suspect hidden characters (zero-width space, BOM) corrupting a file.
Letters and digits: filters to only Unicode letters and numbers. Best for traditional letter-frequency analysis (cryptanalysis, language identification).
Words: splits on whitespace and counts whole words. Best for writing analysis and stylistic checking.
What 'case sensitive' does
Off (default): 'A' and 'a' count together. Best for letter-frequency on natural text where case is incidental.
On: 'A' and 'a' count separately. Useful when case is meaningful — programming identifiers, branded terms, or analyzing capitalization patterns. Note: case-insensitive folding uses the locale's lowercase rules; for most languages this is the conventional Unicode case folding.
Frequently asked questions
›Does it work for Japanese, Chinese, Korean text?
Yes. Letter mode treats each ideograph as one 'letter', so you get hanzi/kanji frequency. Word mode splits on whitespace, which means CJK text without spaces shows as one giant word — use letter mode for those.
›What's the most common English letter?
'E' (about 12.7%), then T (9.1%), A (8.2%), O (7.5%), I (7.0%), N (6.7%). Knowing this is the foundation of breaking simple substitution ciphers.
›Are emoji counted?
Yes in 'all characters' mode. Letter mode filters them out (they're not letters per Unicode classification).
›Why are emoji sometimes split into multiple characters?
Some emoji are multiple Unicode code points (e.g., flags = two regional indicator letters). The counter follows JavaScript string iteration which respects code points but not all grapheme clusters. For most analysis this is fine.
›Can I export the table?
Not yet — copy-paste the rendered table for now. CSV export is on the roadmap.
›How many entries does it show?
Top 50 in the table. The tail count is summarized at the bottom.
›Why don't case-insensitive Greek/Turkish results match my expectation?
Some languages have unusual case rules (Turkish dotted/dotless I; German ß ↔ SS). We use JavaScript's toLowerCase() which follows the default Unicode case folding — usually fine but can surprise in edge cases.
›Does the data leave my browser?
No. All counting runs locally.
Related tools
Last updated: