Guides
Each tool on this site returns a number quickly: a Reading Ease score, a grade level, a reading time, a density percentage, a segment count. The guides exist for the moment after that, when the question changes from what is the number to where did it come from, and how far can I trust it. Each guide takes one metric apart: the inputs the formula reads, why each input matters, how much the result moves when an input changes, how to read the output, and the situations in which the estimate quietly stops describing your text.
The worked examples in the guides are computed at build time by the same engine that runs in your browser on the tool pages. The pipeline behind that engine — tokenization, sentence segmentation, the Flesch coefficients, the syllable heuristic, stop-word handling, the timing constants — is documented in full on the methodology page, and the guides refer back to it rather than restating it.
Readability
Flesch Reading Ease, explained
The Reading Ease formula reads only two ratios — words per sentence and syllables per word — and weights the second far more heavily than the first. This guide takes the formula apart term by term, scores several short original passages to show how much each term moves the result, and explains why the familiar 0–100 reading of the scale is a convention rather than a limit the formula respects.
Flesch–Kincaid Grade Level
The grade-level formula uses the same two ratios with different coefficients, arranged so the output reads as a US school grade. It shares its provenance with Reading Ease — the 1975 recalibration produced for the US Navy — and shares its blind spots too. The guide covers why a grade can fall below zero or climb past any real classroom, and how to read the band labels the checker attaches without treating them as a verdict on your readers.
How syllables are counted
Syllables are the most influential input to both Flesch formulas and the only one that is estimated rather than counted. The engine counts vowel groups and adjusts for a silent final "e"; this guide shows where that heuristic is right, where it is wrong in each direction, and why the errors partially cancel across a passage while still mattering for a single short sentence. The heuristic is built for English and only English.
Timing
Both timing tools divide a word count by a population average: 238 words per minute for silent reading and 150 for prepared speech, neither a pace to aim for. The gap between them is visible on any text — the engine puts a 44-word, 3-sentence sample paragraph at about 11 sec to read silently and about 18 sec to say aloud.
Reading time and words per minute
Reading time is a division, and the guide is honest about both halves of it: what the word count includes (numbers, URLs, hyphenated compounds) and where the 238 wpm figure comes from — Brysbaert's 2019 meta-analysis of silent-reading studies, an average across populations and genres. It explains why the estimate is more useful for comparing two drafts than for predicting any one reader.
Speaking time and speech length
Speaking time applies the same word count at 150 words per minute, a typical prepared-speech pace. The guide works the arithmetic in both directions — how long a script will run, and how many words fit a fixed slot — and covers what stretches delivery beyond the estimate: figures that count as two words but are voiced as four, and pauses the word count cannot see.
Counting & density
What counts as a word
Most disagreements between counters come from the definition of a word, not from arithmetic. This guide lays out the rule the engine uses — a Unicode-aware run of letters and digits in which internal hyphens and apostrophes do not break the run — and walks through the edge cases that follow from it: "2,500" as two words, a URL as three, an emoji as none. When your count differs from another tool's, this is where to look first.
Keyword density, explained
Density is a word's count divided by all words in the text, stop words included in the base, with no stemming. The guide explains each of those three decisions and how a different base or a stemmer would change the percentage for the same text, and is explicit about what the percentage is not: it describes repetition in the text and nothing beyond it.
Characters
SMS length and character counting
A text message is measured in segments, not characters, and the segment size depends on which characters appear. The guide explains the GSM 03.38 rules the engine follows: 160 septets in a single GSM-7 segment, dropping to 153 per segment once a message concatenates; 70 UTF-16 code units in a single segment, or 67 concatenated, once any character outside the GSM set — most emoji, for example — forces UCS-2 encoding. Its examples show one character tipping a message into a second segment.
Back to the tools
Read a guide, then return to the tool with a sharper question. The text analyzer shows every count at once; the readability checker, the keyword density checker, the word frequency counter, and the speaking time calculator each isolate one metric. When a guide's explanation is not enough — when you want the exact rule, the coefficient, or the test that pins a behaviour — the methodology page is the source both the guides and the tools answer to.
Analysis happens locally in your browser — nothing you paste is transmitted or stored. See the methodology page.