Paste a draft in English, French, Spanish or German. Get a contextual rewrite in the same language, protect your figures and quotes, then compare it with the original.
300 free words a day. No account needed.
Free 300 words a day. 5 EUR for 20,000 words, one-off.
Illustrative example: On 3 March, our company will open a new office in Lisbon. The office will have 12 employees and will mainly handle customer support. Recruitment begins in January on the careers page.
Illustrative example: On 3 March, our company will open a new office in Lisbon. The office will have 12 employees and will mainly handle customer support. Recruitment starts in January on the careers page.
Illustrative example, not a measured result
Choose a light, balanced or strong intensity, paste your text and see what changed. Your original and the new version stay next to each other.
Keep this key to restore your credits in another browser.
Three facts, sourced from Anthropic's help center and the press coverage of August 11 to 13, 2026.
On August 11, 2026 Anthropic said it watermarks text generated by Claude. The mark is statistical (the same family as Google's SynthID-Text): it lives in the choice of words, not in hidden characters, so you cannot see it or strip it with find and replace.
How the watermark worksIt applies worldwide to Claude models released since August 2, 2026, in the app and through the API. It survives copy-paste and light edits, and is only weakly detectable on very short texts.
Which texts are affectedThe timing follows the EU AI Act transparency obligations (article 50) that started to apply in August 2026. A side effect: a human text you proofread or translate with Claude comes back watermarked and can be flagged as AI-written.
Why false positives happenThree simple steps from a stiff draft to a text you are happy to sign.
Add between 10 and 200 words in French or English. Protect any phrase you want to keep exactly as written.
Go light for a gentle polish, balanced for a clear refresh or strong for a deeper reworking. Meaning stays at the center.
Read the result next to your original, check the highlighted changes and copy the version you like best.
Professional writing where the draft needs your tone, not a generic one.
Turn stiff drafts into messages that sound like you wrote them.
02Refresh descriptions while keeping specs, prices and brand tone.
03Tighten phrasing and rhythm without losing your argument or sources.
04Edit client drafts faster, then compare each version before delivery.
What the official documentation says on 11 October 2026.
Updated · 11/10/2026
As of 11 October 2026, Anthropic says new models launched in the EU since 2 August 2026, including Sonnet 5.5, are marked globally on supported surfaces, not only for EU users. Rewrites and translations may reduce detectability. Absence of a mark does not prove human origin.
Anthropic: How Claude marks AI-generated contentAs of 11 October 2026, OpenAI offers text verification on request to approved organizations. Its public provenance check documentation covers images and audio. It does not show that all ChatGPT text is marked, and a not_detected result does not prove human authorship.
OpenAI: Content provenanceUnmarker rewrites with context in mind, so sentences read smoothly and still say what you meant.
Phrases are reshaped in context, so your text stops sounding like a template and starts sounding like a person.
Sentences flow the way people actually write, in French or in English, without awkward word swaps.
See the original and the new version together. View the final text only, highlight added or moved words, or open the full view with deleted and added portions.
Numbers, URLs, quotes and the phrases you protect are preserved. Paragraphs and lists keep their structure.
Show or hide figures calculated from your source and result: renewed wording, lexical variety, word counts and rewritten passages.
Variants are prepared using AI, then the final wording is assembled automatically. You always make the final call on the text you publish.
Editorial rewriting gives you control. Marker removal is a different goal.
You decide the strength, protect critical details and see each change. The result is a text you can review, adjust and own.
Rewriting may reduce some signals but cannot guarantee removal, and the rewrite may itself create a marker. This tool cannot certify removal, and there is no web plagiarism search.
Turn rough AI-assisted notes into a clear text that carries your own ideas and your own voice.
Refresh a draft so it reads warmly and keeps the angle you chose for your readers.
Give product pages, newsletters and articles a more distinctive tone that is pleasant to read.
Polish reports, emails and summaries so they sound like you wrote them on a good day.
The more words you buy, the lower the price per word. No subscription.
€0.25 per 1,000 words
€0.20 per 1,000 words
€0.18 per 1,000 words
€0.15 per 1,000 words
Free trial: 300 words per day, up to 3 tries, with 10 to 200 words per text.
Informational guides to help you understand how AI text, watermarks and verification are discussed today.
An overview of what people mean by removing a Claude watermark, and what is actually known about it.
Learn moreGuideA guide to the idea of detecting a Claude watermark and how official verification is handled.
Learn moreGuideWhat to know about ChatGPT watermarks, how they are discussed and where the facts stand.
Learn moreGuideAn introduction to SynthID-Text, the text watermarking approach and what it is designed to do.
Learn moreGuideA practical look at cleaning AI text: stray symbols, odd spacing and formatting leftovers.
Learn moreGuideHow to make Claude drafts read more naturally, with tips on voice, rhythm and wording.
Learn moreGuideA primer on AI watermark detection, its limits and the tools people talk about.
Learn moreGuideWhat a Claude AI detector can and cannot tell you, explained in plain language.
Learn moreGuideYour starting point for every guide about AI watermark removal and related topics.
Learn moreSeven free cleaners run directly in your browser, so your text stays on your device. They are separate from the rewrite tool.
Find and remove zero-width characters, odd spaces, soft hyphens and lookalike letters.
Open the toolFree, in-browserReplace or remove em dashes with commas, periods, colons or parentheses. Diff included.
Open the toolFree, in-browserOne pass: hidden characters, Markdown, smart quotes, dashes, spacing. Pick what to fix.
Open the toolFree, in-browserCount AI writing tells by category: vocabulary, connectors, patterns, punctuation, rhythm.
Open the toolFree, in-browserEvery character with its code point, name, category, block and invisible or lookalike flag.
Open the toolFree, in-browserConvert curly quotes and apostrophes to straight ones, or back.
Open the toolFree, in-browserTurn Markdown into plain text: headings, bold, links, lists, tables, code fences.
Open the toolAccording to Anthropic's help page, new models launched in the EU since 2 August 2026, including Sonnet 5.5, are marked globally on supported surfaces. Here is what that means for your text, and how a careful editorial rewrite helps you keep your own voice.
02Raw AI drafts often read smooth but generic. Rewriting them to sound like a specific voice is genuine editing work. Claiming to erase a provider mark is something else, and this tool neither promises it nor can prove it.
03As of October 11, 2026, Anthropic and OpenAI do not offer the same visibility into text marking. Here is what each signal means, who can check it, and how contextual rewriting helps you keep a text clear, accurate and in your own voice.
No. The rewrite stays in the language you submit. English, French, Spanish and German are supported, and there is no automatic translation.
Provider markers are technical signals, while editing changes wording, structure and rhythm. We focus on editorial rewriting. We offer no detector and no guaranteed marker removal, and absence of a mark never proves human origin.
Open the original and result side by side in three diff views. You can also enable optional lexical metrics, which count real word-level changes instead of estimating them.
Your source text is sent to Anthropic Sonnet 5.5 for rewriting. Results expire within 24 hours. Failed jobs refund the words used.
Paste your text, choose an intensity and compare the result in seconds. Your first 300 words today are free.
300 free words a day. No account needed.