Only selected code points: zero-width spaces, BOMs, unusual spaces, direction controls, or other flagged characters. Every removal appears in the before-and-after view.
Selective Unicode cleaner
Claude Watermark Remover & Unicode Cleaner
Remove observable hidden characters from pasted text without silently normalizing or rewriting visible content. This independent tool does not claim to detect or erase Claude’s model-level watermark.
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.
Removal boundary
Clean exact characters; treat statistical signals separately
Character cleanup is directly verifiable in the diff. Claude’s model-level watermark is a different mechanism, so this page does not turn Unicode removal into a claim about statistical detection.
A clean Unicode report does not mean a Claude statistical watermark is absent. Anthropic’s official detector and technical documentation are not yet public.
What a Claude watermark remover can remove exactly
This cleaner works on the observable layer: exact characters present in your pasted text. It identifies zero-width characters, byte order marks, unusual spaces, bidirectional controls, and other formatting code points with their line and column. You decide which types to remove, and the diff exposes every change.
Character cleaning is not statistical watermark removal
Anthropic says supported Claude models weave an imperceptible watermark into generated text at the model level. It can travel with copied text and may persist through editing. Removing hidden Unicode does not prove that this statistical signal is gone, because the two mechanisms are different.
When to use the remover, cleaner, or rewrite workflow
- Use the cleaner when you need to inspect clipboard artifacts, invisible formatting, or unexpected characters.
- Use the diff when rendering, parsing, accessibility, or publishing behavior matters.
- Treat rewriting separately because it changes visible content and may alter facts, tone, citations, or code behavior.
For the first-party technical boundary, read Anthropic’s current marking guidance.
Clear answers, narrow claims
Frequently asked questions
Can this tool remove Claude’s model-level watermark?
It can remove only the hidden Unicode and formatting artifacts you select. Anthropic describes its text watermark as model-level, and its technical detection mechanism is not public, so this tool cannot verify or guarantee removal of that statistical mark.
What does the Claude watermark cleaner remove?
It reports exact Unicode code points such as zero-width characters, byte order marks, unusual spaces, and direction controls. Nothing is removed until you select a character type and review the result.
Does cleaning change the visible text?
The cleaner does not normalize, paraphrase, or replace visible words. Some invisible characters can still affect layout, emoji, scripts, or source code, so review the diff before using the output.
Does paraphrasing guarantee watermark removal?
No. Anthropic says heavily edited, paraphrased, translated, or mixed text may lack a detectable mark. That is a stated detection limitation, not a guarantee for any particular rewrite.