Advanced Text Analytics & Word Counter
Count words, grade readability, spot weak sentences and clean up text — entirely inside your browser, in nine languages.
Text tools
Hold Alt for the second versionWith nothing selected these apply to the whole document. Select a paragraph to narrow them down.
Ctrl+Z and Ctrl+Shift+Z undo and redo every tool, Ctrl+F opens find & replace, Ctrl+S saves a .txt, Ctrl+Alt+plus and minus change the text size.
How it works
Write, paste or open
Type straight in, paste a draft, or drop a .txt, .md, .docx or .odt file. A file never overwrites what you already have: it waits until you say replace or append.
Set it up
Pick a target length, set your own reading speed, force the language if the guess is off, or switch the live analysis off entirely and stay fully manual.
Read the marks
Long sentences, passive voice, adverbs, hedges and padding are painted onto the text itself. Click any count to jump to the first one.
Clean up and move on
Fix it with tools that respect your selection and undo cleanly, then export a report — or send the text to another oLoveTools tool without downloading it.
Real word boundaries
Words and sentences come from Intl.Segmenter, the ICU segmenter your browser already ships. It knows "don't" is one word and that $3.50 is not two sentences — and it segments Japanese and Chinese properly instead of counting glyphs.
Seven readability formulas
Flesch, Flesch–Kincaid, ARI, Gunning Fog, Coleman–Liau, SMOG and LIX, plus Fernández Huerta for Spanish and Portuguese. The one shown first is the one that suits the language you actually wrote in.
Sentence-level style check
Every sentence is graded on its own, so the tool can point at the paragraph that drags instead of handing you one number. Passive voice, adverbs, filler words and long-winded vocabulary are marked in place.
Keywords, pairs and phrases
Density for single words, two-word pairs and three-word phrases, with stop words filtered using the list of the detected language rather than one bag of nine languages mixed together.
An editor that forgives
Undo and redo across every tool, a hold-to-compare button that shows the text before the last change, find & replace with regular expressions, and tools that only touch what you selected.
Chained to the rest of the suite
Send the finished text straight to the Markdown previewer, the text-to-speech tool, the diff viewer or the code-image maker. No download, no re-upload, no round trip through your Downloads folder.
The analysis runs off the main thread
Segmenting a long document, grading every sentence and building three n-gram tables costs real CPU time. WordFlow does all of it inside a Web Worker, so the caret never stutters while you type — and past 60,000 characters it stops running on its own and waits for you to press the button.
Nothing leaves the tab
There is no upload, no analytics call carrying your text, and no server round trip — the files you open are read with the browser's own FileReader. The local draft option is off by default; when you switch it on, the text is kept in this browser's localStorage and nowhere else, and switching it off deletes it.
A word counter that reads what you wrote
Bloggers checking a word target, copywriters trimming a meta description to 155 characters, students hitting an essay length, SEO specialists reading keyword density, and editors hunting the passive voice — all without pasting a client's draft into somebody else's server.
Length, reading time, keyword density, readability and prose quality — computed in your tab, in nine languages, with nothing sent anywhere.
Frequently Asked Questions
QIs my text sent to a server?+
No. Every count, formula, word list and regular expression runs in JavaScript inside your tab, including the files you open — they are read with the browser's FileReader and never uploaded. You can disconnect from the internet and the tool keeps working.
QHow are words and sentences counted?+
With Intl.Segmenter, the Unicode segmenter built into your browser. Splitting on spaces gets "don't" wrong and splitting on full stops turns $3.50 into two sentences; more importantly, Japanese and Chinese have no spaces at all, so counting one word per character — which is what most counters do — is simply wrong. Where Intl.Segmenter is unavailable, a Unicode-aware fallback takes over.
QWhich readability formula should I trust?+
The one shown first is chosen for the detected language: Fernández Huerta for Spanish and Portuguese, Flesch Reading Ease for English, French and German, and LIX for Russian, Hindi, Japanese and Chinese, where formulas built on English syllables mean nothing. All seven are listed underneath so you can compare them.
QIs the sentiment score AI?+
No, and it is not labelled as such. It is a per-language dictionary of scored words, extended with negation handling so "not great" is not read as praise, and with intensifiers so "extremely good" outweighs "fairly good". A dictionary is honest about being a dictionary; treat it as a rough signal, not a verdict.
QWhat does the style check actually look for?+
Sentences over 25 words (and over 40, more strongly), passive constructions, adverbs, filler and hedging words, long-winded vocabulary that has a plainer alternative, the same content word repeated inside a short window, and stray double or trailing spaces. Each one is highlighted in the text so you can see where it is.
QCan it open Word documents?+
Yes. A .docx or .odt file is a zip archive with an XML document inside, so WordFlow unpacks it and pulls the text out in your browser. Plain text, Markdown, CSV, HTML, subtitles and RTF are read directly. Nothing is uploaded to convert it.
QWill I lose my text if I reload?+
By default, yes — nothing is stored. If you want a safety net, switch on "Keep a local draft" and the text is written to this browser's localStorage as you type. It stays on this device, it is never sent anywhere, and turning the option off deletes it immediately.