diff --git a/readme.md b/readme.md new file mode 100644 index 0000000..9eab43b --- /dev/null +++ b/readme.md @@ -0,0 +1,48 @@ +
![]()
Hide encrypted text inside zero-width Unicode characters. Send secret messages through any text field — chat, email, documents — completely invisible to the naked eye.
+ ++Firefox Add-on · Chrome Web Store (coming soon) +
+ +--- + +## ZWUS Encoding + +Text is encoded using **ZWUS (Zero Width Unicode Standard)** — each character's code point is converted to a target base, then each digit is mapped to a designated zero-width Unicode character. + +| Base | Characters Used | Unicode Points | Notes | +|------|:-:|---|---| +| **ZWUS-3** | 3 + 1 separator | U+180E · U+200B · U+200D · sep U+00AD | Most compatible — fewest characters, safest across platforms | +| **ZWUS-6** | 6 + 1 separator | U+200D · U+200F · U+00AD · U+2060 · U+200B · U+200E · sep U+200C | Balanced | +| **ZWUS-8** | 8 + 1 separator | U+200D · U+200F · U+00AD · U+2060 · U+200B · U+200E · U+180E · U+FEFF · sep U+200C | Smallest output but uses characters some platforms strip | + +**Compatibility:** ZWUS-3 > ZWUS-6 > ZWUS-8 +**Output size:** ZWUS-8 < ZWUS-6 < ZWUS-3 + +Some platforms actively strip or collapse certain zero-width characters. If your encoded text breaks on a platform, drop to a lower base. + +## Encryption + +### PLAIN +No encryption. Encodes text directly. + +### SPECK32/64 ECB +Block cipher in ECB mode. Each character is encrypted independently — identical plaintext characters produce identical ciphertext. Useful when you want varied-looking output from the same input, but **not secure under real threat models** since patterns in the plaintext leak through. + +### SPECK48/96 CTR +Stream cipher mode with a random 24-bit nonce prepended to the output. Identical plaintexts produce different ciphertexts each time. Significantly more secure than ECB. **Rotate your password after roughly every 100 encryptions** to avoid nonce reuse risk within the 24-bit space. + +## How ZWUS Works + +1. Take each character's Unicode code point. +2. Convert it to the chosen base (3, 6, or 8). +3. Map each resulting digit to its assigned zero-width character from the alphabet. +4. Join digits together; separate characters with the base's designated separator (also zero-width). + +The result is a string of invisible Unicode that carries the full original text. Decoding reverses the mapping: split on the separator, look up each zero-width character to recover the digit, parse the base-N number back to a code point, and reconstruct the string. + +## Use ZWUS in Your Own Projects + +The encoder/decoder is a standalone npm package: **[zwus on npm](https://www.npmjs.com/package/zwus)**