Substitution Cipher Solver & Encoder

In a simple substitution cipher every letter is replaced by another fixed letter, like the cryptograms in newspapers. With 26! possible keys it cannot be brute-forced, but it can be solved by search: this tool starts from many keys and keeps swapping letters while the result looks more like English, judged by how common its four-letter groups (quadgrams) are. It also encodes and decodes with a known keyword or cipher alphabet.

How to use this tool

1. Paste the ciphertext; at least 80 letters works best.
2. Leave the mode on Solve and press Calculate (it takes about a second).
3. If a few letters look wrong, add them as hints such as Q=E and solve again, or try another seed.
4. To encode or decode with a known key, switch the mode and enter the keyword or full alphabet.

Formula

score(key) = Σ log₁₀ P(quadgram) over every four-letter window of the decrypted text

The first start matches letters by frequency (the commonest cipher letter becomes E, and so on); later starts make five random swaps in the best key so far, and every sixth start is fully random. Two letters are swapped at a time and the swap is kept if the score rises; after 1,500 swaps without improvement every pair is tried once more until none helps. The highest-scoring key over all starts wins. Probabilities come from about 1.2 million letters of modern English web prose; the 3,000 most common quadgrams are stored, and others get a low floor value.

Example

The default ciphertext begins "Vtvzb lozcys nwv zctvz…". With 30 restarts and seed 1 the solver recovers "Every spring the river behind our village rises above its banks…" — all 305 letters — and the key A→Q, B→K, C→U … that was used to make it. With the keyword ZEBRAS, "FLEE AT ONCE. WE ARE DISCOVERED!" encodes to SIAA ZQ LKBA. VA ZOA RFPBLUAOAR!