About

Data leakage prevention requires rigorous input sanitization. This tool automates the obfuscation of sensitive string literals (PII) before data enters public channels. Unlike simple find-and-replace, it employs algorithmic masking to preserve document structure while mathematically destroying the semantic value of targeted segments.

When masking data, the information entropy H of the string is effectively increased for the viewer, rendering the original signal x irretrievable without the key. This tool handles various obfuscation vectors: Unicode block replacement (Redaction), chaotic diacritic injection (Zalgo), and structural scrambling.

Risk Advisory: Incomplete redaction (e.g., blurring without sufficient radius) is reversible via deconvolution algorithms. For high-security contexts, always use character replacement (Total Masking) rather than visual obfuscation.

Formulas

The Zalgo corruption intensity I follows a summation of random combining marks:

C_out = C_in + ceil(I)∑k=0 rand(U_comb)

Where U_comb represents the set of Unicode combining diacritics in range 0x0300 to 0x036F.

For PII masking, the function M applies condition P:

{

█ if test(P, x) is TRUEx otherwise

Reference Data

Method	Logic Symbol	Description	Reversibility
Block Redaction	S → █	Replaces all characters with Full Block (U+2588).	FALSE (Irreversible)
Symbol Mask	S → *	Replaces characters with fixed symbols (*, #, ?).	FALSE
Zalgo / Glitch	S + U_comb	Injects Unicode combining marks.	TRUE (Sanitizable)
Scramble	perm(S)	Permutes internal characters (Typoglycemia).	PARTIAL
Hashing	sha256(S)	Cryptographic digest of the string.	FALSE (One-way)
Blur (CSS)	filter	Visual blur overlay (Text remains in DOM).	TRUE (Insecure)

Frequently Asked Questions

Yes, if using the "Block" or "Symbol" modes. These modes replace the underlying ASCII/Unicode values entirely. However, if you use "Blur" or "Zalgo", the original text may still be recoverable or selectable in the DOM. Always verify the output before publishing.

Yes. You can select "Custom Regex" to define specific patterns (e.g., social security numbers). The tool uses the JavaScript RegExp engine locally in your browser.

No. The information is mathematically destroyed. The replacement function is non-bijective, meaning multiple inputs map to the same output (the block character), making inverse calculation impossible.

The tool uses the HTML5 Canvas API to render the censored text as a raster image. This creates a "flattened" PNG file where the text is pixels, not selectable characters, adding an extra layer of security.