Guides

Every counter on WordCntr.com follows a stated rule, and every rule has consequences that are not obvious from the number on screen: a hyphen joins two words into one; a period after "Dr" starts a new sentence. These guides explain those mechanisms one unit at a time — what the engine does, where its convention differs from another tool's, and how far to trust the result — so that a disagreement with another tool has a reason you can find.

Each guide runs short original texts through the same engine the tools use, at build time. The methodology page's worked paragraph — "Dr. Smith arrived at 9 a.m. She left early." — comes out as 10 words, 43 characters, 3 sentences and 1 paragraph, and the sentence figure is one more than a reader would count. Each guide answers that kind of why.

Counting

What counts as a word

The rule is one sentence long — an unbroken run of letters, digits, apostrophes and hyphens — and nearly every disagreement between word counters follows from it. The consequences: hyphenated compounds and contractions count once, a decimal number splits at its period, a pasted URL contributes a token for every segment, and a hyphen standing alone as a list bullet registers as a word. The guide also covers what the unique-word figure merges.

Sentences, paragraphs and lines

These units depend on punctuation and line breaks rather than on the words themselves. Run "Ferries run hourly.↵Last one at dusk.↵↵Book ahead." through the engine and it reports 3 sentences, 2 paragraphs and 4 lines, 1 of them blank: a single line break separates lines but not paragraphs, only a blank line does that, and a line break alone never ends a sentence. The guide covers the abbreviation over-count, the word-processor paragraph mark, and the trailing-newline rule that keeps the line count in step with wc -l.

Characters & letters

Characters, code points and emoji

The character counter tallies Unicode code points. "Well done 👍" is 11 characters; add a skin-tone modifier — "Well done 👍🏽" — and it becomes 12, because the modifier is its own code point riding on the same picture. Flags, joined family emoji, line breaks and trailing spaces follow from the same rule. The guide also explains why some platforms report a different figure: they count UTF-16 code units, under which most emoji cost two.

Letters and Unicode normalization

The letter counter counts only code points Unicode classifies as letters, with digits, spaces and punctuation tallied separately. It normalizes text to NFC first, so a composed accented letter and the same letter typed as a base letter plus a combining mark agree: "café" is 4 raw code points and its decomposed twin is 5, yet the letter counter finds 4 letters in the first and 4 in the second. The guide explains the normalization step, why the two pages' character totals can differ, and the honest edge: a combining mark with no composed form is counted under punctuation and symbols.

Timing

Reading time and words per minute

Reading time is the word count divided by 238 words per minute — the average adult silent-reading rate for non-fiction in Brysbaert's 2019 meta-analysis — so a 1,000-word piece is estimated at 4 min 12 sec. That is an estimate of the middle of a wide range, not a prediction for any particular reader. The guide explains the evidence behind the constant, how the word rule's quirks leak into the estimate, and when to round up or down.

Speaking time and speech length

Speaking time divides by 150 words per minute, a common prepared-delivery pace, so the same 1,000 words take about 6 min 40 sec. The guide works the arithmetic in both directions — how long a draft runs, and how many words fit a time slot — and states the limits: the constant is a population average for prepared speech, nothing beyond the word count enters the calculation, and the result is a planning figure, not a promise.

Transforms & tools

Case conversion rules

The case converter offers 5 transforms — UPPERCASE, lowercase, Title Case, Sentence case, aLtErNaTiNg — each with a stated rule. Title Case is the simple convention: the first letter of every word token capitalized, the rest lowercased, no minor-word list, so "3rd edition" becomes "3rd Edition" with the digit-led word left alone. Sentence case inherits the sentence rule's abbreviation limitation, aLtErNaTiNg alternates per letter starting lowercase, and every transform uses the default Unicode English mappings — which the guide explains is wrong for Turkish text.

How the word finder matches letters to words

The word finder answers one question — which dictionary words can be assembled from the letters you hold — by letter-multiset containment: a word is makeable when it needs no more copies of any letter than your rack supplies, and each blank ("?") covers one missing copy. Racks run from 2 to 20 tiles with at most 2 blanks; results are grouped by length, longest first, then alphabetically. The guide also states the dictionary's provenance: the public-domain ENABLE list of 172,820 entries, which predates the 2006 additions — so "qi" and "za" are not in it.

From the guides to the tools

The word counter reports words, characters, sentences and paragraphs, and the dedicated character, letter, sentence, paragraph and line counters, the reading time calculator, the case converter and the word finder each link back to the guide that explains their numbers. When a guide is more than you need, the methodology page states every rule in one place, with the known over-counts and the input limit.

Counting happens locally in your browser — nothing you type is transmitted or stored. Every number on this page is computed at build time by the same engines the tools run; conventions are named where they apply. See the methodology page.