TextAnalyzer ProPaste a text and read it the way an editor would: counts and reading time, six readability formulas with a consensus grade, the tone and the words that set it, keywords and a word cloud, sentence structure with passive voice and long sentences, writing suggestions, and two texts compared side by side. Everything stays in your browser: nothing is uploaded.
The text is remembered in this browser until you clear it. Select part of it to analyse that part only. Text files up to 2 MB can be opened or dropped.
Readability appears once there are ten words
The scores need at least 10 words and one sentence, and settle down past a hundred words.
-
Enter text to analyse
-
U.S. school grade
-
Years of education
-
Letters, not syllables
-
Polysyllabic words
-
Characters per word
Neutral
Score +0.00 on a scale from -1 to +1
The score is positive minus negative, divided by their total. A text with three positive words and one negative scores +0.5; a text with none scores 0 and reads neutral. "Not good" counts as negative, "not bad" as positive.
Words of three letters or more, ignoring case and about 230 common words (the, and, of, is, very ...).
One to three percent marks a clear topic word; above three percent a reader notices the repetition.
The forty most frequent words, sized by how often they occur, appear here.
A readable text mixes lengths; an average between 15 and 20 words with a few short ones for emphasis is the usual advice.
Press Show highlights to read the text with its long sentences marked. Hover a sentence for its word count.
A line of each sentence's length appears here once there are two sentences; the dashed lines mark 25 and 35 words.
Type or paste into the Text box, press Paste to take the clipboard, drop a text file on the box or open one with the button, or press Sample text to see every tab filled. The analysis runs as you type; the status pill in the toolbar shows the word count, the reading level and the tone at a glance. Select part of the text and press Analyse the selection to look at one passage.
Counts gives words, characters, sentences, paragraphs, unique words, syllables, the longest word and the reading and speaking time. Reading Level shows the consensus grade on a scale, the level it corresponds to and the audience it suits. Writing Suggestions lists what an editor would mark first; Reading Time by Paragraph shows where the long stretches are.
The Readability tab shows Flesch Reading Ease and the five grade formulas, each with its formula behind the ⓘ mark and a note on what drove it, then the five grades side by side with their average. Formulas that agree mean a reliable estimate; a formula that stands out tells you whether long words or long sentences are the cause.
Tone places the text on a scale from very negative to very positive and lists the exact words that were counted, so you can see whether "critical" meant a complaint or a review. Keywords shows the most frequent meaningful words with their density, the two-word phrases that repeat, and a word cloud you can download as a picture.
Structure counts paragraphs and sentences, shows how the sentence lengths are spread, how many sentences look passive, how many connectors and -ly adverbs there are, and highlights the long sentences in the text itself. The difficulty line shows where the text gets heavy; Typical Targets compares your length and grade with blog posts, news, marketing copy, documentation and papers; Plain-English Alternatives points at words with a simpler twin.
Set as baseline keeps this version's numbers so that every edit shows as a change against it. Compare puts two texts side by side in one table. Save analysis keeps a snapshot in this browser; the Saved window lists them with Load and Delete and, from two saves on, the trend of word count and grade. Export gives JSON, CSV, a text or Markdown report, a copied summary or a printed page.
Readability tools often show one score and call it the truth. TextAnalyzer Pro shows Flesch Reading Ease, Flesch-Kincaid, Gunning Fog, Coleman-Liau, SMOG and the Automated Readability Index with their formulas on the page, a consensus grade, and a note when the sample is too small for a formula to be trusted. The textbook example scores exactly what the textbook says.
Every judgement comes with the evidence: the words that made the tone positive or negative, the sentences that look passive, the long sentences highlighted in the text, the adverbs and connectors it counted, the plain-English alternatives it found. Where a count is a heuristic, the page says so, so you can overrule it.
Set a baseline, then rewrite: the Overview shows every number against the version you started from. Put two drafts in Compare and see which one reads easier, in one table. Save a snapshot after each pass and watch the word count and grade move in the Saved window's trends.
A manuscript nobody has read, a contract, a cover letter, a patient leaflet before it is approved: the analysis runs in your browser and nothing is sent to a server. What you type is kept in this browser's storage so it is there tomorrow, and Clear or Settings removes it. The only copies that leave are the files you download.
The analysis runs as you type, with a short pause on texts past twenty thousand characters so the box stays responsive. A 10,000-word chapter is analysed in a fraction of a second; the word cloud and the difficulty line are drawn when their tab is opened.
JSON with every number, keyword and sentence length for a spreadsheet or a script; CSV of the metrics and keywords; a plain-text or Markdown report for an email or a wiki; a one-click copy of the summary; or a printed page of the whole analysis, which your browser can save as a PDF.
All six formulas were worked out by testing readers on passages and fitting a line to two things that are easy to count: how long the sentences are and how long the words are. Flesch Reading Ease (1948) gives a score out of 100; Flesch-Kincaid (1975) turns the same two numbers into a U.S. school grade for the Navy's manuals. Gunning Fog (1952) and SMOG (1969) count the share of words with three or more syllables; Coleman-Liau (1975) and the Automated Readability Index (1967) use letters per word, because typewriters could count letters but not syllables. None of them reads the text. A clear sentence full of long technical words scores badly; a vague sentence of short words scores well. That is why the consensus matters more than any single score, and why the tool shows the words and sentences behind the numbers.
The tone score counts words from two lists, about 220 positive and 230 negative, flips a word when it follows a negation, and reports positive minus negative over their total. It is quick, transparent and language-specific, and it fails in predictable ways: it does not know that "the critical reviews were glowing" is good news, that "sick" can be praise, or that a sentence can be sarcastic. It is useful as a check on drift, when a product description has quietly filled with "problem", "risk" and "fail", or when a complaint letter reads friendlier than you intended, and the list of counted words lets you judge each case. Treat it as a reading aid, not a verdict.
The keyword list is a frequency count with about 230 function words removed (the, and, of, is, very and the rest), so what remains is what the text is about. Density is a word's share of all words: one to three percent marks a clear topic word, above three percent readers notice the repetition, and search engines have long since stopped rewarding it. Two-word phrases catch the terms a single-word count breaks apart, such as "social media" or "data security". The word cloud sizes the forty most frequent words by their count; it is a picture of emphasis, not of meaning, and its best use is spotting the word you did not know you were leaning on.
Flesch Reading Ease runs from 0 to 100, higher being easier: 60 to 70 is plain English, 90 and above is very easy, under 30 is academic. The other five give a U.S. school grade: Flesch-Kincaid and Automated Readability from sentence length and word length, Gunning Fog and SMOG from the share of long words, Coleman-Liau from letters per word. The consensus grade is the average of the five, and the reading level names the audience it suits.
The formulas divide by the number of sentences and words, so a sentence or two produces wild numbers. Ten words and one sentence is the minimum; the scores settle down past a hundred words, and SMOG is designed for thirty sentences or more, so it is marked approximate below that.
Every word is checked against a list of about 220 positive and 230 negative words; a negation in the two words before (not, never, isn't) flips it. The score is positive words minus negative words, divided by their total, so it runs from -1 to +1, and the label follows the score. It is a lexicon count, not an understanding of the text: sarcasm and context escape it, which is why the words it counted are shown so you can judge.
A sentence ends at a full stop, question mark or exclamation mark followed by a space, or at a line break; decimals such as 3.5 and common abbreviations such as Dr., e.g. and U.S. do not end one. Words are runs of letters and digits in any script, with apostrophes kept inside them. Paragraphs are separated by an empty line.
Nowhere. The analysis runs in your browser and nothing is sent to a server. The text box, the two compare boxes, the saved analyses, the baseline and the reading speed are kept in this browser's storage so they are there when you come back; Clear, the Delete buttons and Clear data in Settings remove them. There is no account.
The counts, keywords, phrases, word cloud, sentence lengths and comparison work for any language written with spaces between words. The readability formulas, the syllable estimate, the sentiment lexicon, passive voice, transitions, adverbs and the plain-English list are built for English and give meaningless numbers for other languages.
It is free, with no registration and no limits. Like every page on ZeroDataUpload, it is supported by advertising.