Word Counter

Text length becomes a practical constraint in applications, essays, captions, scripts, transcripts, and publishing systems. A word total answers whether a draft fits, while sentence length, repeated phrases, and readability estimates point to different editing problems.

A “word” is not a universal unit. Apostrophes can stay inside a token, decimal punctuation can keep a number together, and hyphens can divide text into separate tokens. Case and accent rules also decide whether forms such as Draft and draft, or café and cafe, belong to the same frequency group. Comparisons are meaningful only when the same rules are used for both drafts.

Readability scores reduce a passage to sentence length and estimated syllables. They can reveal dense prose, but they cannot judge accuracy, organization, tone, accessibility, or whether specialist vocabulary is necessary. Repetition counts have a similar limit: a frequent term may be needless filler, or it may be the subject that keeps the passage coherent.

• Word and character counts support length limits and rough timing, but another system may tokenize the same text differently.
• Unique-word share describes vocabulary variety within one draft; it does not measure quality or originality.
• Frequency and phrase lists find repeated wording without deciding whether the repetition is useful or excessive.
• Flesch scores flag long sentences and syllable-heavy words without predicting comprehension for a specific audience.

How to Use This Tool:

Start with the counting rules that match the destination, then keep them unchanged while comparing revisions.

1. Paste a draft into Text, drop one TXT file, or browse for it. The analysis updates as the text changes and accepts up to 200,000 characters.
2. Decide whether numeric tokens count, whether uppercase and lowercase forms group together, whether accents fold, and the shortest token to include. A high Minimum word length can leave no countable words.
3. Use Advanced when repetition is the main concern. Common-word filtering affects the frequency, phrase, and chart views only; it does not redefine the headline word total or readability scores.
4. Read Text metrics first, then use Word frequency and Repeated phrases to locate wording worth reviewing. Keep phrase boundaries inside sentences unless cross-sentence sequences are genuinely useful.

Interpreting Results:

The word total reflects the current number and minimum-length settings. Unique words, top-word share, and phrase rankings also depend on case and accent grouping. When common English words are hidden, percentages in the frequency table use the remaining repetition tokens rather than the full word total.

• Easy read means a Flesch Reading Ease score of 80 or higher.
• Plain read covers scores from 60 up to, but not including, 80.
• Dense read covers scores from 30 up to, but not including, 60.
• Very dense means a score below 30.

Treat those labels as editing signals. Short lists, headings, abbreviations, technical terms, and non-English text can make the score less representative of real reading difficulty.

Technical Details:

Words are matched as runs of Unicode letters and combining marks, with an internal straight or curly apostrophe allowed. Integer and decimal-like numeric tokens are optional. Hyphens and most other punctuation divide tokens. Sentence boundaries use terminal periods, exclamation marks, and question marks; paragraphs are separated by blank lines.

Formula Core:

Flesch Reading Ease combines average sentence length with estimated syllables per word. Higher values usually indicate shorter sentences and fewer syllables.

Flesch-Kincaid Grade Level uses the same counts on a grade-scale output.

Here W is the countable word total, S is the sentence count with a minimum of one, and Y is the estimated syllable total. Lexical density is the unique-word count divided by the word count, multiplied by 100.

Syllables are estimated with an English-oriented vowel-group heuristic after accent folding. The displayed averages, percentages, and readability scores are rounded to two decimal places; the underlying calculations keep full numeric precision. Reading time rounds upward from 200 words per minute, and speaking time rounds upward from 125 words per minute, with a minimum display of one minute.

Transformation Core:

Frequency rows count normalized tokens and rank them by count, then alphabetically for ties. Repeated phrases are contiguous one-, two-, or three-token windows. They restart at sentence boundaries by default. The row limit of 5 to 50 changes only how many ranked rows appear, not the complete text metrics.

Privacy Notes:

Pasted text and TXT files are read and analyzed in the browser. The source is not uploaded for counting. Local processing is still not a reason to leave confidential drafts unattended on a shared device or to include sensitive text in an exported report.

Worked Examples:

References:

• Unicode Standard Annex #29: Unicode Text Segmentation, Unicode Consortium, Unicode 17.0.
• Measuring Readability, Indian Health Service.