Nothing is selected by default. Low-risk artifacts are separated from characters that can affect emoji, bidirectional text, line wrapping, or complex scripts.
Text-first cleanup
Claude Text Watermark Remover
Paste prose, notes, or drafted copy to inspect invisible characters and selectively remove only the types you choose. Visible wording is not rewritten by the cleaner.
Analyze + Clean stay local Nothing pre-selected No fake detection score
Exact Unicode inspection
Analyze first. Clean only what you select.
Analyze + Clean stay local
- 1AnalyzeFind exact UnicodeLocal
- 2CleanRemove selected typesLocal
- 3RewriteChange visible contentOff
Scope: Claude’s model-level watermark is not hidden Unicode and is not detected by this inspector.
0 characters
Inspection report
Characters found
0flagged characters
0character types
0lines checked
No inspected artifacts found
This checks known invisible Unicode and formatting controls only. It is not a detector for Claude’s model-level statistical watermark.
0 types selected
Selective Cleaner result
Review every change
Only the exact code points you selected were removed. No normalization, rewriting, or other replacement was applied.
BeforeRemoved markers
AfterCleaned output
Step 3 · Optional server processing
Rewrite is unavailable
Unlike Analyze and Clean, a configured rewrite would send your content to a model provider. No provider or server endpoint exists in this MVP.
Rewriting changes visible text or code. It may reduce a detectable statistical signal, but this tool cannot verify or guarantee that outcome.
Text editing boundary
Preserve wording during cleanup; review rewriting separately
Selective cleanup removes only chosen Unicode types. Rewriting changes visible language and needs editorial review; neither action can prove that Anthropic’s model-level mark is absent.
Substantive paraphrasing may reduce a detectable statistical signal, but it can change meaning and cannot guarantee watermark removal. No model provider is configured in this MVP.
Text cleanup that preserves visible wording
This page is designed for prose, notes, documentation, and publishing copy. The local inspector separates observable Unicode artifacts from visible language, so you can remove selected characters without automatically rewriting the draft.
Why selective cleanup matters for text
Not every invisible character is unwanted. Non-breaking spaces affect wrapping, direction controls affect bidirectional text, and joiners or variation selectors can affect scripts and emoji. Nothing is selected by default, and higher-risk findings are marked for review.
When text rewriting becomes a different task
Anthropic says heavily edited, paraphrased, translated, or mixed text may not retain a detectable mark. A rewrite changes visible language and can introduce meaning, tone, or citation errors. It should therefore be reviewed as an editorial change—not presented as guaranteed watermark removal.
Clear answers, narrow claims
Frequently asked questions
Can I clean Claude text without changing the wording?
Yes, when you use selective character cleanup. The cleaner removes only the exact code point types you choose and does not paraphrase visible words.
Which invisible characters need extra review?
Direction controls, zero-width joiners, variation selectors, and some spaces can affect emoji, multilingual scripts, display order, or wrapping. They are flagged for review rather than silently selected.
Is the pasted text uploaded?
Current Analyze and Clean operations run in the browser. Rewrite is unavailable because no model provider or server endpoint is configured.
Will cleaned text pass an AI detector?
This tool makes no such promise. Unicode cleanup does not remove model-level statistical signals, and general AI detectors are not the same as Anthropic’s forthcoming mark detection mechanism.