A rewording tool restates your text in different words. Ours does that and adds the thing most such tools leave out: a highlight of exactly which words changed, plus an AI score before and after from our own detector. Three a day, up to 260 words each, with no account.
The more useful half of this page is about when rewording is the wrong goal. It is a narrow operation, and people reach for it to solve three problems it does not solve. Knowing which is which saves more time than any tool does.
Narrower than the name suggests, which is the root of most disappointment with these tools.
It substitutes vocabulary and reshapes phrasing while trying to hold the meaning still. That is the whole operation. It works at the level of words and clauses, not at the level of argument or structure.
So it will reliably change how a sentence reads. It will not reorder your paragraphs, strengthen a weak point, notice that two sections contradict each other, or tell you that paragraph four is unnecessary. Those are editing decisions, and they are the ones that make writing better.
Two different things get sold under the same name. Word-level work swaps individual words for synonyms and leaves the sentence shape alone, which is what most basic tools do and what produces the characteristic thesaurus feel: technically different, slightly wrong, often less clear than the original. Sentence-level work rebuilds the sentence, which reads better and moves further from your original phrasing.
Ours works at the sentence level, which is why the diff matters. The further a tool moves from your wording, the more opportunity it has to move your meaning with it, and you cannot audit that in prose you have only read for flow.
In order of how often people arrive wanting exactly this.
1. Plagiarism. Putting a copied passage in different words does not make it yours. It is the same idea in different words with the same source uncited, which is what plagiarism is. Changing the vocabulary changes whether a string-matching checker spots it, and not whether you have attributed someone else's work. The fix is a citation. Paraphrasing properly is a different skill and involves attribution.
2. A weak argument. If a paragraph is unconvincing, it is usually missing evidence, not better adjectives. You will get a more fluent version of the same gap. This is the most common waste of time with these tools: polishing a sentence that should have been deleted or supported.
3. An accusation. If someone has flagged your work as AI-generated, running it through this is actively harmful. Your version history is what evidences authorship, and a reworded draft replaces that record with a single block of changed text. Gather evidence instead.
Four habits, the first of which is the only one that is non-negotiable.
Read the diff, not the prose. The output will read fine. That is what it is optimised for. The question is whether anything moved that should not have: a number, a date, a hedge, a claim. In our own testing one pass returned the sentence "Roughly a year into it" when nothing in the source mentioned any timescale. An invented specific is the real failure mode, and it reads smoothly.
Work a paragraph at a time. Partly the 260-word anonymous cap, mostly that you will genuinely review a paragraph and will not genuinely review two pages.
Keep your technical terms. If a word is a term of art, a synonym is wrong rather than equivalent. Lighter settings leave vocabulary alone, which is what you want on anything specialist. Light varies rhythm and cuts filler and does not touch your vocabulary; Balanced reshapes phrasing more broadly; Maximum replaces most of the wording, is a paid mode, and is the most likely to substitute a term you needed.
Work on your own writing, not a source. See wrong reason 1.
The new passage, the changed words highlighted, and an AI score before and after from our own detector, labelled as ours. What nobody can give you is another checker's verdict: Turnitin and the rest do not expose theirs to third-party tools, so a published pass rate is a number the vendor chose. Current per-request sizes are on the plans page.
Yes, with no account: three rewordings a day at up to 260 words each, resetting daily. You get the full reworded passage, a highlight of which words changed, and an AI score before and after from our own detector. The highlight is the part worth using, because it is how you check that a number or a claim did not move while the prose was being improved.
In practice very little; the names are used interchangeably. The distinction worth drawing is between word-level and sentence-level work. Word-level tools swap individual words for synonyms and leave the sentence shape alone, which produces the thesaurus feel: technically different, often less clear. Sentence-level tools rebuild the sentence, which reads better and moves further from your original, so the diff matters more.
No. Rewording a copied passage leaves you with the same idea in different words and the same source uncited, which is what plagiarism is. It may change whether a string-matching checker flags the text, which is not the same thing as having attributed someone else work. The fix is a citation, not a tool.
Yes, and it is the main thing to watch rather than a footnote. The output reads fluently because fluency is what it optimises for, so a softened hedge or a strengthened claim is easy to miss. It can also insert something that was not there: in our testing a pass added a timescale the source never mentioned. Read the highlighted changes against your original, especially around numbers, dates and citations.
No, and this is where lighter settings matter. If a word is a term of art, a synonym is not equivalent to it, it is wrong, and a reader in your field will notice immediately. The lightest setting varies sentence rhythm and cuts filler while leaving your vocabulary alone, which is the right choice for specialist writing. Heavier settings replace more wording and are correspondingly more likely to substitute a term you needed.
Two hundred and sixty per request on the free anonymous tier, three times a day. That is a per-request ceiling rather than a daily pool, so it is three short passages rather than one long one, which also happens to be the right working scale: you will review a paragraph properly and you will not review two pages properly. Current per-request sizes for each plan are on the plans page.
Which of the two categories you actually need, and why.
See the comparison →The version that involves attribution, which a tool cannot do for you.
Read the guide →How to audit a rewrite instead of trusting it.
See the method →Three a day, 260 words, with the changed words shown.
Reword a paragraph →Three a day, no account, with every changed word highlighted so you can see what moved.