Autokey Cipher Decoder

Decrypt autokey cipher messages with known keys or use cryptanalysis tools to break unknown keys. Includes dictionary attack and brute force capabilities.

0 letters
-
- / 4 threads

How Autokey Decryption Works

Decrypting an autokey cipher with a known seed key is straightforward but has an important characteristic: each recovered plaintext letter immediately becomes part of the keystream for decrypting the next character. This creates a cascade effect where a single error propagates through the entire remaining message.

With seed "KEY" and ciphertext "KXRAVD":

Position123456
CiphertextKXRAVD
KeystreamKEYA*T*T*
PlaintextATTACK

Positions marked with * show recovered plaintext letters being fed back into the keystream. If position 1 were decrypted incorrectly, every subsequent position would also be wrong.

Decrypting Without the Key

Breaking the autokey cipher requires different techniques than standard Vigenere cryptanalysis, because the key never repeats. Kasiski examination and index of coincidence analysis are ineffective here.

Dictionary Attack

The most powerful approach exploits a unique vulnerability: natural language plaintext appears directly in the keystream after the seed. Testing common English words at various positions often produces readable text, and each correct guess cascades forward to reveal more.

For example, if "THE" appears at position 5 of the plaintext, it also appears at positions 5-7 of the keystream. Testing this hypothesis against the ciphertext either produces sensible output at subsequent positions (confirming the guess) or gibberish (ruling it out).

Probable Word Analysis

If you suspect the message contains specific words (names, dates, technical terms), test them at various positions. Each match reveals adjacent keystream characters, which can be tested further. Even a 3-character match often cascades to reveal 8-10 additional characters.

Statistical Fitness Scoring

Modern automated analysis scores candidate decryptions against expected English letter frequencies, bigram frequencies, and common word patterns. Messages longer than 100-150 characters provide enough statistical weight for reliable automated breaking.

Brute Force with Estimated Key Length

If you know or can estimate the seed length, systematically test all possible seed values. The decoder ranks results by linguistic fitness scores, presenting the most probable solutions first. This approach is feasible for seeds up to about 6-8 characters.

Using This Decoder

  1. Enter the ciphertext in the input field
  2. Choose your mode: known-key decryption, dictionary attack, or brute force
  3. Set parameters such as estimated seed length or probable words
  4. Review ranked results sorted by confidence score

The decoder visualizes the cascade effect during decryption, showing how each recovered plaintext letter feeds into the keystream for subsequent positions. This educational feature helps you understand why autokey decryption errors propagate and why the self-keying mechanism creates both strengths and weaknesses.

Frequently Asked Questions

How long should ciphertext be to break the autokey cipher?

For reliable automated decryption, aim for at least 100-150 characters. Messages over 200 characters give high success rates with dictionary attacks. Very short messages (under 50 characters) usually require additional hints or context to break.

Why do errors cascade during autokey decryption?

Each decrypted character becomes part of the keystream for the next character. A wrong decryption produces a wrong key letter, which produces another wrong plaintext letter, which creates another wrong key letter, and so on. This is fundamentally different from Vigenere decryption, where errors affect only individual positions.

What makes autokey vulnerable to dictionary attacks?

After the seed keyword, the keystream consists entirely of plaintext. Since plaintext is natural language, it contains predictable words and patterns. Testing common words at various keystream positions exploits this directly, making autokey uniquely susceptible to this type of attack.

Can known plaintext break the autokey cipher?

Yes, and very effectively. Knowing any portion of the plaintext lets you reconstruct the keystream from that point forward, cascading to reveal the rest of the message and even the original seed keyword. This makes autokey particularly weak against known-plaintext attacks.