Advanced Text Statistics

Beyond counting words: what "statistics" really means here

A plain word counter answers one question — how long is this? This tool answers a dozen more. Alongside the basic counts it computes vocabulary richness, average word and sentence length, the longest and shortest word and sentence, a syllable total, a readability score, and two frequency analyses: the top content words and the distribution of every letter. The goal is not just to measure your text but to characterize it — to show whether your prose is repetitive, dense, or varied — which is what makes it a writing-improvement tool rather than a tally.

Vocabulary richness, the metric most counters skip

Richness is the ratio of unique words to total words, shown as a percentage. A passage that reads "the dog saw the dog and the dog ran" has 8 words but only 5 unique ones, giving 62.5%. The figure is a quick proxy for repetition: marketing copy and children's books score low because they reuse key terms deliberately, while a varied essay scores higher. Watch the trend, not the absolute number — richness naturally falls as a document grows longer, because common words like "the" and "and" keep recurring, so compare passages of similar length.

The readability score, and how syllables are estimated

The tool computes the Flesch Reading Ease score from three measured quantities — words, sentences, and syllables — using the standard formula 206.835 − 1.015 × (words ÷ sentences) − 84.6 × (syllables ÷ words). Higher is easier: 90+ is very easy (around 5th grade), 60–70 is plain English most adults read comfortably, and below 30 reads like a professional or academic document. Long sentences and long words both drag the score down. Syllables are estimated with a fast heuristic: each word is lowercased, a silent trailing e is dropped, and runs of vowels are counted as one syllable each. It is not perfect on unusual words, but because the formula averages across your whole text, a few miscounts barely move the result.

Top words and letter frequency

The top-10 content words list deliberately filters out a built-in set of common stop words — the, and, is, you, and roughly a hundred others — so it surfaces the terms that actually carry your meaning rather than grammatical filler. This is genuinely useful for checking SEO focus or spotting an overused crutch word. The letter-frequency bars rank a–z by count; in ordinary English text, e, t, a, and o dominate, so a wildly different profile is a hint that your text is code, a list, or another language.

A worked example

Paste the two sentences: The fox runs fast. It jumps high! The tool reports 7 words, 7 unique (100% richness), 2 sentences, an average sentence length of 3.5 words, and roughly 8 syllables. With such short, simple sentences the Flesch score lands very high — in the "very easy" band — exactly as you would expect for one-syllable words in four-word sentences. The top-words panel, after removing stop words like the and it, would surface fox, runs, fast, jumps, and high.

Reading and speaking time

Two time estimates round out the report: reading time at 238 words per minute (a well-cited silent-reading average) and speaking time at 130 words per minute (an unhurried presentation pace). Short snippets show in seconds, longer pieces in minutes. Treat both as planning figures — comprehension speed swings with topic and formatting — but the speaking figure is reliably useful for sizing a script: a five-minute talk is roughly 650 words.

Use cases and tips

• Tightening prose. A low Flesch score plus a long "longest sentence" tells you exactly where to break a run-on or swap jargon for plain words.
• Spotting overuse. If a single content word tops the frequency list by a wide margin, you are leaning on it; vary your phrasing.
• Checking SEO focus. The top-words list should feature your intended keyword — if it does not, the page may be off-topic.
• Sizing speeches. Use speaking time, not reading time, to fit a talk into its slot before you rehearse.
• Mind the limits. The readability model assumes English; scores on code, URLs, or other languages are unreliable, and sentence detection depends on ., !, and ?, so unpunctuated text skews the per-sentence figures.

Everything is computed in your browser as you type, so even a confidential draft never leaves your device — it works fully offline.