Nine free writing tools in one place. Check your text for AI, humanize an AI draft, proofread it, count the words, measure readability, and check it against the web for plagiarism. No account, no sign-up. Paste once below and every reading updates as you type.
Each one is a page you can paste into and get an answer from immediately. No tool asks for an email, holds a result behind a signup, or counts your checks.
Paste a draft and see which sentences an AI checker keys on, one by one, before anyone else runs it.
The same reading with its four measurements written out: tell density, rhythm uniformity, stock lexicon, hedging.
Rewrite the flagged clauses in your own measured style, one at a time, with you approving every change.
Misspellings, confused words, broken agreement and stray commas, marked where they sit with the correction one click away.
Delve, leverage, foster, robust, seamless. Every stock word marked, with a replacement and a reading of your vocabulary range underneath.
Words, characters with and without spaces, sentences, paragraphs, syllables, reading time and speaking time, on every keystroke.
Flesch Reading Ease, Flesch–Kincaid, Gunning fog and SMOG on one paste, plus the sentences running long enough to lose a reader.
Reads a document against itself: repeated passages, quotations with nobody attached, citations missing a reference entry, figures with no source.
Write in the box and it records the session — time at work, how much you cut, any large paste — then signs it so anyone can check.
Every pattern the engine looks for, written out with an example, the reason it reads as machine-made, and the fix.
Write with every tell underlined in place, watch the meters move as you revise, and export a signed record of the drafting.
Check the signature on an evidence report against the public key the report carries. Open to anyone, no account.
The readings at the top of this page answer different questions, and only one of them is usually the one you came for. Find yours here and go straight to it.
A draft can be flawless on grammar and still read as machine-made, because the two engines measure unrelated things. Sentence rhythm is what a detector keys on. Spelling is what a reader trips over.
Every number on this page came out of code that shipped with the page. That is why the results are instant, why there is nothing to charge for a check, and why the tools keep working when you switch the network off.
41 surface patterns and 3 whole-document cadence detectors, each explained in one sentence.
51 of them carry a mechanical fix, so the correction is one click rather than a lecture.
Behind the spelling track, generated from validated bases with irregular verb and plural tables.
Flesch Reading Ease, Flesch–Kincaid, Gunning fog and SMOG, each with the word class it is defined on.
Every tool here has a real edge, and knowing where it is makes the readings more useful rather than less.
It searches your text against itself, not against the web, a journal database, or an archive of submitted work. It finds repeats, unattributed quotations, citation problems and unsourced figures. It cannot tell you whether a passage exists somewhere else, and it is not a substitute for an institutional check.
It reports how machine-made a text reads, computed from rhythm, lexicon, hedging and tell density. Who typed it is a separate question, and no tool that reads finished text can answer it. Commercial detectors run their own classifiers on thresholds they do not publish, so they will not always agree with this reading.
Flesch and its relatives measure syllables per word and words per sentence. A short, simple sentence that says nothing scores well. Use the grade band as a signal about pitch, then read the flagged long sentences yourself.
Written for the situations the tools do not fix by themselves: an accusation, a class policy, a detector that disagrees with a student who did the work.
No. There is no sign-up, no login, no quota and no billing anywhere on this site. Every tool on this page is JavaScript that runs in the tab you are reading, so there is no per-check cost to meter.
No limit. There are two floors, and they exist because the arithmetic below them is meaningless rather than because anything is being withheld: the detectability score starts at 25 words, and the readability formulas start at 100. Counts, proofreading and the originality check work on a single sentence.
Nothing is uploaded in order to check it, and nothing is sent to a detector. The one thing that persists is the editor: opening a draft there keeps it in your own browser storage so a refresh does not lose your work, and clearing it removes it.
Yes. Load the page once and disconnect. The rules, the lexicon and the formulas are already in the page, so every reading keeps computing with the network off. That is also the simplest way to confirm the first two answers for yourself.
This check reads your document against itself. It does not search the web, any journal database, or any archive of previously submitted work, so it cannot tell you whether a passage appears somewhere else. It finds text repeated inside the document, quotations with no attribution nearby, citation problems, and figures with no source. It is not a substitute for an institutional plagiarism check such as Turnitin.
No tool can promise that, and this one does not. What you get is the specific list of constructions producing the machine-made reading in your draft, and a decision on each one. What a commercial detector reports on a given day is its own call, on a classifier it retrains whenever it likes.
Paste into the box at the top of this page. It runs all four engines at once and tells you which reading is the one worth acting on, then links you to the tool that explains it.
Free, no account, and the whole thing runs in your browser.