Sentence Cipher

A multi-language library (Go & TypeScript) that encodes arbitrary binary data into natural-looking English sentences. Hide your secret messages in plain sight!

Overview

Sentence Cipher transforms any byte sequence into grammatically correct English sentences using a workplace/office theme vocabulary. Unlike simple substitution ciphers, it constructs meaningful sentence structures that mimic professional communication.

Example

Mode: String
Input:  "Hello"
Output: "ruth trains isabella prints. carl cleans daily."

Mode: Natural
Input:  "Hello"
Output:
"""
Subject: Code Review

Hi Team,

Ian imports sandra os. Specifically, Leo sorts daily. 

Best regards,
Samantha
"""

Mode: Binary
Input:  "Hello"
Output: "ruth trains isabella prints. carl cleans daily."

How It Works

The library offers three primary modes of operation to suit different needs:

📝 String Mode

Standard encoding that transforms text into office-themed sentences. It handles UTF-8 strings directly, making it perfect for short text messages or chat applications.

📧 Natural Mode

Advanced encoding that structures data into a natural-looking email format. It wraps the encoded sentences with realistic:

• Subjects: "Code Review", "Meeting Notes", "Project Update"
• Openers: "Hi Team,", "Good morning,"
• Connectors: "Specifically,", "Additionally,"
• Closers: "Best regards,", "Cheers,"

This mode generates output that looks indistinguishable from a real workplace email, providing better cover for your data.

💾 Binary Mode

Raw byte encoding for any file type (images, documents, executables). It maintains data integrity by treating the input as a raw byte stream. The output looks the same as String Mode but ensures that binary data is perfectly preserved during the round-trip.

---

Encoding Scheme

The cipher uses three word lists, each containing exactly 256 words (mapping to byte values 0-255):

Word List	Purpose	Examples
Names	Subjects & Indirect Objects	adam, alex, alice...
Verbs	Actions	helps, assists, supports...
Objects	Direct Objects	reports, documents, files...

Sentence Patterns

Data is encoded using three different sentence patterns depending on the remaining bytes:

1. Full Sentence (3 bytes): [Subject] [Verb] [IndirectObject] [Object].
2. Short Sentence (2 bytes): [Subject] [Verb] daily.
3. Minimal Sentence (1 byte): [Subject] works.

Installation

Go

go get github.com/kittizz/sentence-cipher

Node.js / TypeScript

npm install sentence-cipher
# or
pnpm add sentence-cipher

Usage

Go Library

package main

import (
    "fmt"
    sentencecipher "github.com/kittizz/sentence-cipher"
)

func main() {
    // 1. Basic encoding/decoding
    message := "Hello World"
    encoded := sentencecipher.EncodeString(message)
    fmt.Println(encoded) 
    // Output: "ruth trains isabella prints. carl cleans daily..."

    // 2. Natural Mode (Email style)
    naturalEncoded := sentencecipher.EncodeNatural([]byte(message))
    fmt.Println(naturalEncoded)
    // Output: "Subject: Project Update\nHi Team,\nRuth trains..."

    // 3. With Encryption Key
    cipher := sentencecipher.NewCipher("my-secret-key")
    secureEncoded := cipher.EncodeString(message)
    
    // Decode
    decoded, _ := cipher.DecodeString(secureEncoded)
    fmt.Println(decoded)
}

Node.js / TypeScript Library

The JavaScript library provides a Type-Safe API with full support for all modes.

import { createCipher, createDefaultCipher } from 'sentence-cipher';

// Initialize
const cipher = createDefaultCipher();

// --- String Mode ---
const encoded = cipher.encodeString("Hello World");
const decoded = cipher.decodeString(encoded);
console.log(encoded); 
// "ruth trains isabella prints. carl cleans daily."


// --- Natural Mode ---
// Natural mode works with Uint8Array to support any data
const input = new TextEncoder().encode("Hello World");
const naturalEncoded = cipher.encodeNatural(input);

console.log(naturalEncoded);
/*
Subject: Team Sync
Hi Everyone,
Ruth trains isabella prints. Carl cleans daily.
Best,
*/

const naturalDecoded = cipher.decodeNatural(naturalEncoded);
console.log(new TextDecoder().decode(naturalDecoded)); // "Hello World"


// --- Binary Mode ---
// Perfect for files or raw data
const data = new Uint8Array([72, 101, 108, 108, 111]); // "Hello" bytes
const binaryEncoded = cipher.encode(data);
const binaryDecoded = cipher.decode(binaryEncoded);


// --- Encryption Key ---
// Provide a key to shuffle the word lists deterministically
const secureCipher = createCipher("my-super-secret-key");
const secureMsg = secureCipher.encodeString("Secret Data");

// Must use the same key to decode
const decryptedMsg = secureCipher.decodeString(secureMsg);

Command Line Tool (CLI)

You can also use the Go CLI to encode/decode files or text from the terminal.

# Install
go install github.com/kittizz/sentence-cipher/cmd/sentencecipher@latest

# Usage
sentencecipher "Hello World"
sentencecipher -d "ruth trains isabella prints..."
sentencecipher -n "Generate natural email"
sentencecipher -k "my-key" "Encrypted message"

Technical Details

Key Derivation & Security

When a key is provided, the library uses SHA-256 to hash the key. The hash is used to seed a Fisher-Yates shuffle algorithm, which randomizes the order of the Names, Verbs, and Objects word lists. This ensures that without the correct key, the sentence mapping is completely different, effectively encrypting the message.

Expansion Ratio

The encoding transforms binary data into English text, which naturally increases the size.

• Expansion: Approximately 15-20x original size.
• Efficiency:
  • 3 bytes → ~4 words
  • 2 bytes → ~3 words
  • 1 byte → ~2 words

License

MIT License